| ||||||||||||||
|
|
|||||||||||||



From the *Department of Anesthesiology and Intensive Care, Lund University Hospital, Sweden;
Department of Anesthesiology, Malmö University Hospital, Sweden;
Department of Pharmaceutical Biosciences, Uppsala University, Sweden; and
Department of Oncology, Lund University Hospital, Sweden.
Address correspondence and reprint requests to Åsa Rudin, Department of Anesthesiology and Intensive Care, Lund University Hospital S-221 85 Lund, Sweden. Address e-mail to asa.rudin{at}skane.se.
| Abstract |
|---|
|
|
|---|
METHODS: In the present study we investigated morphine pharmacokinetics in patients who had undergone liver resection (n = 15) compared to a control group undergoing colon resection (n = 15). Morphine was administered IV by patient-controlled analgesia. Plasma concentrations of morphine, morphine-6-glucuronide, and morphine-3-glucuronide were measured 2–3 times daily for the first two postoperative days. Pain intensity scores were assessed three times daily and respiratory rate and sedation scores every third hour.
RESULTS: There were no differences in morphine requirements 1.1 (0.8–2.5 [median, interquartile range]) mg/h (liver resection) and 1.5 (1.1–1.7) mg/h (colon resection) [P = 0.84]) or in pain intensity scores (P > 0.3) between the groups. Plasma morphine concentrations were higher in patients undergoing liver resection than in the control group (P < 0.01) reflecting a lower rate of morphine metabolism. Plasma morphine concentrations were correlated with the volume of liver resection (P < 0.02). However, plasma concentrations of morphine-6-glucuronide and morphine-3-glucuronide did not differ between the groups (P = 0.62 and P = 0.48, respectively). There was a higher incidence of sedation (P = 0.02), but not respiratory depression (P = 0.48), after liver resection.
CONCLUSION: The study demonstrates that plasma concentrations of morphine are higher in patients undergoing liver resection compared with patients undergoing colon resection. Sedation scores were higher in patients undergoing liver resection. Caution is therefore recommended when administering morphine to this patient group.
| Introduction |
|---|
|
|
|---|
Morphine is mainly conjugated in the liver and the metabolites morphine-6-glucuronide (M6G) and morphine-3-glucuronide (M3G) are excreted via the kidneys and bile (4). Renal impairment therefore results in higher plasma levels of morphine glucuronides compared to morphine (5,6). Similarly, morphine pharmacokinetics is affected by reduced hepatic function (6–9). Although slowly developing hepatic disease, such as liver cirrhosis and hepatic carcinoma, may induce a compensatory extra-hepatic metabolism, the pharmacokinetic consequences after an acute reduction in hepatic function after liver resection may be different. However, morphine pharmacokinetics after liver resection and its clinical consequences for analgesia and adverse effects have not been systematically studied.
The aim of the present, prospective, clinical study was to evaluate morphine pharmacokinetics and adverse effects in patients undergoing liver resection compared to a control group undergoing colon resection.
| METHODS |
|---|
|
|
|---|
Midazolam 2.5–7.5 mg was given orally for premedication and general anesthesia was induced with thiopental and fentanyl. Muscle relaxation was provided with either rocuronium or atracurium. Anesthesia was maintained with nitrous oxide in oxygen, isoflurane or desflurane, and incremental doses of fentanyl and a muscle relaxant.
The liver resection was made via a Mercedes incision, intraoperative ultrasonography was routinely used for tumor localization, and transparenchymal dissection was performed with an ultrasonic dissector. The resected liver volume was estimated by the two surgeons immediately after surgery. In all patients undergoing liver resection cholecystectomy was routinely performed. The colon resection was made via a low midline incision. A PCA pump (CADD-Legacy® 6300 or CADD-Prizm® 6101, Smiths Medical MD, St. Paul, MN) for IV administration of morphine hydrochloride (1 mg/mL; APL, Apoteket AB, Umeå, Sweden) was started at the end of surgery. The continuous infusion rate was 0.5 or 1.0 mg/h with a rescue dose of 0.5–2.0 mg and a lockout interval of 10 min (based on patient age and extent of liver resection).
Supplementary oral doses of acetaminophen were given as needed postoperatively. After a 3–8 h observation period in the postanesthesia care unit the patients were transferred to a surgical ward. An anesthesiologist visited the patients daily for treatment evaluation and dose adjustments. When postoperative analgesia was considered satisfactory with a combination of oral acetaminophen and tramadol, the PCA-pump was disconnected, tramadol was started and the study was completed. The patient's preoperative medication was continued on the second postoperative day (Day 2).
Venous blood samples for morphine, M3G, and M6G were collected two or three times per day during the first two postoperative days. The samples were planned to be collected at steady state concentrations of morphine. The distribution half-time for morphine (t1/2
) is 6 min (10) and sampling was avoided if a bolus dose had been administered within the preceding 20 min. Samples were drawn into heparinized glass vacuum tubes, separated immediately by centrifugation (3000 rpm, 10 min), and stored at –20°C (11). All samples were analyzed for morphine, M3G, and M6G by liquid chromatography followed by tandem mass spectrometry (LC/MS/MS) (12). A volume of 100 µL plasma was used for the assay. After a precipitation step, the samples were desalted on a HyPurity C18 guard column, separated on a ZIC HILIC (hydrophilic interaction chromatography) column and detected with MS/MS. The ranges were 0.8–500, 1.5–1000, and 0.5–500 ng/mL for morphine, M3G, and M6G, respectively.
In the surgical ward, the nurses registered respiratory rate and sedation score every third hour. Pain scores were registered three times per day. After each blood sample, the patient evaluated pain at rest and during mobilization with a verbal numeric rating scale (0–10), and sedation was evaluated with the Observer's Assessment of Alertness/Sedation Scale (13). The scale has five levels (1–5). At Level 1 the patient is asleep and unresponsive to tactile stimulation, and at Level 5 the patient is alert. Sedation was defined as an Observer's Assessment of Alertness/Sedation Scale score
4 and respiratory depression was defined as a respiratory rate <8 per minute. The patients' weights and plasma creatinine concentrations were measured daily.
Data from the PCA pumps (rate, bolus, time for bolus dose, and volume infused) were downloaded (CADD-Diplomat, SIMS Deltec, MN) or printed after the end of the treatment.
Since we had no available data for a formal power analysis, we estimated the minimal relevant difference of M3G/morphine and M6G/morphine ratios between the groups to be 1 standard deviation (sd). For a power of 0.80 (Type II error of 0.20) and a Type I error of 0.05, 15 patients in each group were required (14).
Ratios of M6G/morphine and M3G/morphine were calculated. The initial ratios from the first postoperative day (Day 1) and median values of the ratios for each patient from Day 1 and Day 2 were compared between the groups. Morphine clearance was estimated by dividing the mean morphine dosing rate by the mean morphine steady-state plasma concentration. The mean hourly morphine consumption was calculated for the 3 h preceding each blood sample. The correlation between initial morphine concentration, initial ratios (M6G/morphine and M3G/morphine) on Day 1, and volume of liver resected were analyzed.
Data were analyzed for a normal distribution by the Kolmogorov–Smirnov test, and parametric or nonparametric tests were used where appropriate (SPSS 11.5, Chicago, IL). Nominal data were compared by Fisher's exact test or by the
2 test. For comparisons of nonparametric continuous variables the Mann– Whitney U-test was used for unpaired data, and the Student's t-test for parametric data. Friedman's test was used for comparisons between groups when more than two sets of data were compared within the same treatment group. The association between two continuous variables was analyzed by Spearman's correlation coefficient. Corrections for multiple comparisons were made by the Bonferroni method. Data are expressed as median (interquartile range) unless otherwise indicated. Post hoc power calculations were made (14). P values <0.05 were considered statistically significant.
| RESULTS |
|---|
|
|
|---|
|
No difference in pain intensity scores were observed between patients after liver and colon resection (Table 2). The administered dose of morphine per hour during the treatment period did not differ between the groups 1.1 (0.8–2.5) mg/h and 1.5 [1.1–1.7] mg/h, respectively (P = 0.84, Mann–Whitney test). However, in patients undergoing liver resection the PCA pump was a priori set to deliver a lower continuous rate. These patients administered significantly more PCA morphine than patients undergoing colon resection (Day 1: 27.5 mg vs 10.5 mg, [P = 0.02], Day 2: 7.75 mg vs 2.5 mg per day, [P = 0.01], Mann– Whitney test).
|
The patients undergoing liver resection had significantly higher initial plasma morphine concentrations on Day 1 (Fig. 1a, [P < 0.01]) and the morphine clearance was significantly lower in this group compared to the colon resection group (85 vs 130 L/h, P < 0.01, Mann–Whitney). The initial morphine concentration on Day 1 was correlated with the volume of liver resection (Spearman's rho [
] = 0.60 [P < 0.02], Fig. 1a) but there was no correlation between morphine dose and volume of liver resection. The initial Day 1 concentrations of M6G and M3G did not differ between the groups (P = 0.62 and P = 0.48, respectively) and therefore the ratios M6G/morphine and M3G/morphine were significantly lower after liver resection compared to colon resection (P < 0.001, respectively, Figs. 1b and c). Ratios of M6G/morphine and M3G/morphine were inversely correlated with the volume of liver resection (Spearman's rho = –0.56, [P = 0.03] and –0.52, [P < 0.05], respectively Figs. 1b and c). Sedation at the time of the initial blood sample Day 1 is indicated with arrows (Figs. 1a–c).
|
The median ratios of M6G/morphine (P < 0.01, Mann–Whitney, Bonferroni) and M3G/morphine (P < 0.01, Mann–Whitney, Bonferroni) were significantly lower on Day 1 after liver resection compared to colon resection. The difference between the groups was not present on Day 2 (P = 0.56 and 1.0, respectively, Mann–Whitney, Bonferroni, Figs. 2a and b).
|
Significantly more patients in the liver resection group experienced sedation (8 vs 2 patients, respectively [P = 0.02], Fisher's exact test). The incidence of sedation was significantly higher after resections of 50% or more (n = 6) compared to resections <50% (n = 2) (P = 0.032, Fisher's exact test). Two patients in the liver resection group but none in the control group developed respiratory depression (P = 0.48, Fisher's exact test). In one patient with a 60% liver resection continuous morphine infusion was discontinued and the respiratory rate returned to normal within 1 h. In the second patient with a 70% resection naloxone was administered, the continuous morphine infusion was discontinued, the rescue dose was decreased and the patient was observed in the postanesthesia care unit overnight.
Post hoc power analysis with
= 0.05 demonstrated a power of 0.99 (β = 0.01) based on the initial ratios of M6G/morphine and M3G/morphine. To detect a significant difference in respiratory depression (
= 0.05, β = 0.20) an appropriate sample size was estimated to be 110 patients, i.e., 55 patients in each group.
| DISCUSSION |
|---|
|
|
|---|
The metabolism of morphine is often impaired in patients with liver failure, causing a prolongation of the elimination half-life (7,8,15–17). However, the glucuronidation ratios (M6G/morphine and M3G/morphine) after IV administration seem to be maintained (8). Extra-hepatic clearance of morphine has been described in the gut and kidneys in both cirrhotic patients and in individuals with normal liver function. In cirrhotic patients there may be a compensatory increase in this extra-hepatic metabolism (7,15,16,18). On the other hand, patients undergoing liver resection have an acute deterioration of liver function, and pharmacokinetics may therefore differ when compared to patients with slowly progressing hepatic dysfunction.
The hepatic clearance of a drug is the product of hepatic bloodflow and the hepatic extraction ratio. If the extraction ratio is high, the clearance will depend on the hepatic bloodflow, whereas changes in hepatic enzyme activity will have minimal influence (19). For drugs with a low extraction ratio, only a small fraction of the drug delivered to the liver is removed per unit of time and an excess of drug is available for hepatic elimination. Changes in bloodflow will therefore have a minor effect on hepatic drug clearance.
Morphine has an intermediate to high hepatic extraction, so that its hepatic clearance is dependent on bloodflow (4). Most anesthetics decrease portal bloodflow, primarily due to a decreased cardiac output. In the normal physiological state, changes in portal venous flow induce reciprocal changes in hepatic arterial flow to preserve total liver bloodflow (20). During anesthesia, however, although hepatic arterial bloodflow may increase, this is not sufficient to restore the total hepatic bloodflow to normal values (20). In addition, laparotomy per se may result in a decreased hepatic arterial bloodflow, where presumably upper abdominal surgical procedures decrease the liver bloodflow to a greater extent than in lower abdominal surgery (21). Surgical trauma, rather than anesthesia, has been shown to be the main determinant of the alterations in the splanchnic and liver circulation (22). Surgical stress with the associated catecholamine release and activation of the renin–angiotensin system seems to play an important role in the decreased splanchnic blood volume and flow. However, Sear et al. (23) found no differences in the elimination of morphine in patients undergoing various surgical procedures compared to an awake control group.
Our findings of increased morphine concentrations and lower M6G/morphine and M3G/morphine ratios after liver resection may reflect a lower glucuronidation rate, probably due to a decrease in liver bloodflow. However, an additive effect of the concomitant reduction in enzyme activity, due to a reduction in functional liver parenchyma can obviously not be excluded.
Additional medication with acetaminophen, benzodiazepines, and tricyclic antidepressants may influence the pharmacokinetics of morphine (5,15). In the present study premedication with midazolam and postoperative use of acetaminophen were similar in both groups and are therefore not likely to have influenced the results.
The present study indicates that morphine-based postoperative analgesia after liver resection is associated with a higher risk of sedation, and probably of respiratory depression, compared to a colon resection procedure, although the resection per se may partly explain the higher incidence of sedation, alternatives to systemic morphine analgesia would be interesting. Major liver surgery is associated with postoperative coagulation disturbances (24), which may limit the use of epidural analgesia. De Pietri et al. (25) evaluated single-dose administration of intrathecal morphine as an alternative to epidural catheter-based analgesia in patients scheduled for liver resection, but 92% of patients still required IV PCA morphine during the first 48 h postoperatively; therefore, this method would not seem to decrease the incidence of opioid-related adverse effects. The postoperative coagulation disturbance also restricts the use of the opioid-sparing nonselective nonsteroidal antiinflammatory drugs. Opioids with a comparable clinical duration of action, such as oxycodone and buprenorphine, are also eliminated mainly through hepatic metabolism and therefore do not seem to be obvious alternatives to morphine (6). Fentanyl has no active metabolites, but it has a high hepatic extraction ratio and the hepatic clearance is therefore perfusion dependent (26).
In summary, we studied morphine pharmacokinetics and morphine-related adverse effects in patients undergoing liver resection compared to a control group of patients undergoing colonic resection, during postoperative PCA. After liver resection we observed a significant reduction in M6G/morphine and M3G/morphine ratios and an increase in circulating morphine concentrations, mainly due to a lower morphine clearance. Patients who had undergone liver resection had a higher incidence of sedation and a trend toward an increase in respiratory depression. Therefore caution is recommended when administering morphine to this patient group.
| ACKNOWLEDGMENTS |
|---|
| Footnotes |
|---|
Supported by Region Skåne, Lund, Sweden.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
N. F. Quiney, A. Patel, and W. J. Fawcett Factors Affecting Morphine Metabolism Following Major Liver Resection Anesth. Analg., January 1, 2008; 106(1): 348 - 349. [Full Text] [PDF] |
||||
![]() |
A. Rudin and M. U. Werner Factors Affecting Morphine Metabolism Following Major Liver Resection Anesth. Analg., January 1, 2008; 106(1): 349 - 349. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|