Altered hematologic profiles following donor right hepatectomy and implications for perioperative analgesic management


  • Roman Schumann,

    Corresponding author
    1. Department of Anesthesia, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    • Department of Anesthesia, Tufts–New England Medical Center, Box 298, 750 Washington St., Boston, MA 02111
    Search for more papers by this author
    • Telelphone: 617-636-6044, ext. 9328; FAX: 617-636-8384

  • Luis Zabala,

    1. Department of Anesthesia, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    Search for more papers by this author
  • Michael Angelis,

    1. Department of Surgery, Division of Transplant Surgery, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    Search for more papers by this author
  • Iwona Bonney,

    1. Department of Anesthesia, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    Search for more papers by this author
  • Hocine Tighiouart,

    1. Biostatistics Research Center, Division of Clinical Care Research, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    Search for more papers by this author
  • Daniel B. Carr

    1. Department of Anesthesia, Tufts–New England Medical Center and Tufts University School of Medicine, Boston, MA
    Search for more papers by this author


Living liver donors for adult liver transplant recipients undergo extensive liver resection. Partial donor hepatectomies may alter postoperative drug metabolism and hemostasis; thus, the risks and the benefits of pain management for this unique patient population may need to be reassessed. The safety and efficacy of combined epidural analgesia and field infiltration in our initial living liver donor group are presented. A thoracic epidural catheter was placed before general anesthesia in 2 female and 6 male donors (44.2 ± 11.3 years old, mean ± standard deviation [SD], range 26–56). At the end of surgery, incisions were infiltrated (bupivacaine 0.25%), and an epidural infusion was used (bupivacaine 0.1% + hydromorphone hydrochloride 0.02%). Clinical outcomes were followed for 5 days. The time sequence of pain intensity on a 0–10 visual analog scale clustered into 3 phases, the intensity of which differed significantly from each other (2.2 ± 0.6, 0.69 ± 0.2, and 2.37 ± 0.3 respectively, P = 0.028). Right shoulder pain was observed in 75% of the donors. Sedation, pruritus, and nausea were minimal. Consistently maximal international normalized ratio elevation occurred at 17.6 ± 7 hours postoperatively, then slowly declined. Platelet counts were lowest on day 3. No neurologic injury or local anesthetic toxicity was observed. This 2-site approach provided effective, safe, postoperative analgesia for our donors. Universally, coagulopathy ensued, indicating a potentially increased risk for epidural hemorrhage at epidural catheter removal and mandating close postoperative neurologic and laboratory monitoring. Research is needed to advance the understanding of postoperative coagulopathy and hepatic dysfunction in these donors to further optimize their perioperative management, including that of analgesia. (Liver Transpl 2004;10:363–368.)

Minimization of perioperative pain enhances many surgical outcomes,1 and safe, effective postoperative pain management is a clinical goal, an ethical mandate, and a standard of the Joint Commission on Accreditation of Healthcare Organizations.2 Large liver resections may result in transient metabolic impairment3–6 and temporary disturbances in hemostasis.3, 4, 7–11 Hence, development of an optimal perioperative pain management strategy for donors is a complex clinical challenge. Although epidural pain management has been recommended,12 and used successfully in living liver donor surgery,8, 9, 12 a systematic assessment of its efficacy and safety in living donors for adult liver transplantation, who typically donate approximately 60% of their liver,13, 14 is not available.

Reports in operations other than liver donation suggest that epidural pain management, in combination with local anesthetic field infiltration, offers effective, opioid-sparing analgesia15–17 as an alternative to systemic opioids with their potential hepatotoxicity18, 19 and risk of sedation. In the current case series, we evaluated the efficacy and safety of this 2-site approach and assessed postoperative changes in liver function tests and coagulation. On the basis of these results and the current literature, the complexity of perioperative pain management in this unique patient population is discussed.


VAS, visual analog scale; SD, standard deviation; Plt, platelet count; PT, prothrombin time; INR, international normalized ratio; AST, aspartate aminotransferase; ALT, alanine aminotransferase; PCA, patient controlled analgesia; CNS, central nervous system; PACU, post anesthesia care unit.


After approval by the Human Investigation Review Committee, demographics, surgical, and perioperative outcome data were reviewed for our first 8 consecutive liver donors undergoing right hepatectomy for adult liver transplantation. A thoracic epidural catheter between T6 and T11 was placed preoperatively. General anesthesia was induced with either thiopental sodium or propofol and was maintained with isoflurane in oxygen and air. Muscle paralysis was accomplished with vecuronium bromide, pancuronium bromide, or both. At the end of surgery, before extubation, the incisions were infiltrated with plain bupivacaine 0.25%. A continuous epidural infusion of bupivacaine 0.1% with hydromorphone 0.02% was used postoperatively for pain control. Intraoperative epidural catheter use and systemic opioid administration, postoperative rescue pain medication use, and epidural infusion adjustments, as well as duration of epidural pain management, were recorded. The duration of epidural analgesic infusion was based on clinical pain assessment by the Acute Pain Service in consultation with the surgeon. The required laboratory threshold for catheter removal per Acute Pain Service policy was an international normalized ratio (INR) of 1.3 or less. The intensity of postoperative pain on a 0 to 10 visual analog scale (VAS) and the occurrence of nausea, pruritus, and sedation were monitored for 5 days. Examination of the time course of VAS scores of individual patients revealed 3 distinct phases. These phases were analyzed using the Friedman test for nonparametric repeated measures. A P value of 0.05 or less was considered statistically significant. Values are presented as means ± standard deviation (SD). Concurrently, the following laboratory parameters were tracked for 5 days: platelet count (Plt), prothrombin time (PT), INR, aspartate aminotransferase (AST), and alanine aminotransferase (ALT).


The 2 female and 6 male donors were 26–56 years old (44.2 ± 11.3); their body mass index was 27 ± 5.4 kg/m2. Six of these otherwise healthy 8 patients were relatives of the recipients. Five different anesthesiologists provided anesthesia. Intraoperatively, the epidural catheter was used for bolus injections in 3 donors and for bolus and continuous infusions in 5 donors using lidocaine or bupivacaine. During surgery, systemic analgesia consisted of fentanyl 5.2 ± 2.6 mg/kg, and 1 patient received an additional 5 mg of morphine intravenously. The donated liver mass was 60 ± 3.4% of total organ weight as determined by preoperative computed tomography volumetric measurements and confirmed by the actual graft weight after hepatic resection. The remaining liver mass/donor body weight ratio was 0.83 ± 0.2%.

In all patients, the INR and PT were elevated postoperatively, with a maximal rise at 17.6 ± 7 hours postoperatively and not completely normalizing during the study period (Fig. 1). In 3, 2, and 3 patients respectively, the INR declined to 1.3 or less on postoperative day 2, 3, and 4. Concomitantly, all 8 patients experienced a decrease in Plt and an elevation in AST and ALT (Figs. 2 and 3).

Figure 1.

Perioperative values for the international normalized ratio (INR) in 8 patients. Values in means ± standard deviation (SD). INR reference range: 0.9–1.1. Baseline = preoperatively, 0 = day of operation.

Figure 2.

Perioperative platelet counts in 8 patients. Values in means ± standard deviation (SD). Platelet reference range: 150–400 K/μL. Baseline = preoperatively, 0 = day of operation.

Figure 3.

Perioperative alanine aminotransferase (ALT, solid line) and aspartate aminotransferase (AST, dashed line) values in 8 patients. Values in means ± standard deviation (SD). ALT reference range: 0–25 U/L, AST reference range: 10–42 U/L. Baseline = preoperatively, 0 = day of operation.

Postoperative VAS pain scores evolved in 3 phases (Fig. 4). In phase I, consisting of the first 6 hours postoperatively (arrival in the post anesthesia care unit [PACU] = time 0), all patients received supplemental analgesia for an aggregate pain score of 3.1 ± 1.6 at time 0. This phase included titration of the epidural infusion rate with or without a bolus or the administration of systemic medications. Early postoperative right shoulder pain was documented in 6 of 8 patients (75%). Phase II, beginning 7 hours after surgery until the end of the epidural infusion, was characterized by the lowest pain score of 0.69 ± 0.2. During phase II, the epidural infusion (8.4 ± 2.3 mL/h for 69.1 ± 7.9 hours) required little or no adjustment, and no systemic pain medications were needed to maintain very low pain scores. Phase III commenced 75 ± 22 hours postoperatively on discontinuation of epidural pain management. With this transition to patient-controlled intravenous or oral analgesia, the aggregate pain score increased to 2.37 ± 0.3. A statistically significant difference was found between the aggregate VAS scores of phase II and phase III (P = 0.0082).

Figure 4.

Visual analog scale (VAS) pain intensity scores (0–10 scale) during 3 phases of pain management in 7 patients. Horizontal lines within boxes indicate median values, boxes indicate ranges of scores, and “whiskers” indicate standard deviation (SD). Phase I is the first 6 hours postoperatively; phase II is more than 6 hours postoperatively until end of epidural pain management; phase III is postepidural pain management. *Statistically significant difference between pain scores in phase II and phase III (P = 0.0082).

Early postoperative severe right shoulder pain and inadequate incisional pain relief, despite an epidural infusion rate of 8 mL/h, resulted in a change in pain management strategy in 1 patient, whose VAS scores and side effects were subsequently excluded from the analysis. Her epidural catheter had been placed at the level of T9 and when tested produced an appropriate sensory response. This patient required intravenous opioid boluses followed by patient-controlled analgesia as supplements to her epidural infusion to maintain a VAS score of 3 or less.

This donor was also the only one who had an unacceptable degree of central nervous system depression after systemic opioid supplementation, prompted by her refractory right shoulder pain. Approximately 1 hour after the intravenous administration of a total of 2 mg of hydromorphone to alleviate this pain on the day of surgery, she became increasingly sedated and was given intravenous naloxone. The epidural infusion was subsequently changed to an opioid-free solution to avoid possible respiratory depression caused by simultaneous systemic and neuraxial narcotic administration. Suspecting decreased postoperative hepatic drug clearance in addition to a possibly increased opiate sensitivity in this patient, chloroprocaine hydrochloride, which is less dependent on hepatic metabolism, was substituted for bupivacaine in the epidural medication.

Three patients reported pruritus, and 2 reported nausea. These side effects appeared to be of limited clinical import. Neurologic injury or local anesthetic toxicity did not occur. No patient required transfusions to normalize coagulation before removal of the epidural catheter.


First reported in 1989, living liver donation for pediatric recipients usually involves a left lateral segmentectomy in the adult donor liver. This procedure has since been accepted as safe for the donor, with minimal morbidity and mortality.13 The extension of living donation to adult recipients has become possible by performing a right hepatectomy, resulting in a large graft able to meet the metabolic demands of the adult recipient. Hence, living liver donors for this recipient population submit to removal of a larger proportion of their liver, frequently corresponding to 60% of its native volume and sometimes reaching 70%.13, 14 It is recommended that clinicians aim for a graft-to-recipient body weight ratio of at least 0.8% to avoid posttransplantation liver graft dysfunction secondary to a size mismatch between graft and recipient, that is, the “small-for-size” problem.13, 20 Although the minimal necessary size of the remaining liver to ascertain donor safety perioperatively has not been determined, Fan et al. suggested that 30% of the original liver volume suffices to ensure donor survival.14 This issue becomes even more important if complications in the donor arise postoperatively, augmenting hepatic metabolic and synthetic demands. It is unclear to what extent a possible inadequacy of remaining liver size may have affected the recently reported donor mortality.21 In our study, none of the donors (remnant liver to donor body weight ratio of 0.83 ± 0.2%) progressed to postoperative liver failure.

Large hepatic resections can lead to alterations in hemostasis and, possibly, drug metabolism. A transient postoperative coagulopathy in liver donors has been described previously8, 9, 11, 14 and was confirmed in our series, but its pathophysiology is still poorly understood. A direct relationship appears to exist between the prolongation of the prothrombin time and the extent of the liver resection.3, 4, 7, 8 Less characterized is a possible effect of the hepatectomy on the metabolic competence of the remnant liver postoperatively.3–6 Exquisite, acute, and unexpected sensitivity of the central nervous system to the depressant effects of systemic opioid administration was observed in 1 patient in this series and a few additional donors in our institution since, possibly reflecting reduced hepatic opioid metabolism. It is conceivable that a clinically relevant transient alteration in metabolic liver function occurs concurrent with the observed coagulopathy, which could influence pharmacokinetics and dynamics of drugs administered during this time. Even for epidurally administered drugs such as lidocaine, such an impairment could increase their toxicity. Investigations of continuous epidural lidocaine administration during extended hepatectomy have shown that lidocaine blood levels are significantly closer to the toxic range than in controls. These elevations in serum lidocaine levels are related to the extent of the liver resection rather than the diagnosis for which hepatectomy was performed.5, 6 The serum half-life of lidocaine appears to be prolonged in these patients.5 Hence, pharmacokinetic studies in this donor population may contribute to the tailoring of anesthetic and perioperative pain management for this unique patient population. These results also suggest that careful consideration of the most appropriate level of thoracic epidural catheter placement is important to ascertain the minimal epidural infusion rate needed for good pain control. Anesthesiologists caring for these patients face the dilemma that altered drug metabolism may lead to excessive sedation and possibly hepatotoxicity after systemic administration of postoperative opioids,18, 19 whereas a coagulopathy carries the risk of epidural hemorrhage and its potentially devastating sequelae when epidural catheters are repositioned or removed postoperatively.

In our donors with successful epidural pain management, no serious adverse events were observed. However, the small sample size requires cautious interpretation of this finding because serious adverse consequences of epidural analgesia are rare. Our 2-site approach resulted in excellent pain control based on VAS pain scores. When we examined VAS scores across time, 3 phases became apparent, indicating that improvement of pain control should focus on the transition from the operating room to the immediate postoperative period and from epidural to systemic pain management. These “analgesic gaps” may be a result of the epidural management strategy in our institution only. However, the experience of others suggests that such gaps are frequent during transitions from more invasive to less invasive pain control modalities.22, 23

Initial postoperative right shoulder pain was frequent but resolved over time. Diaphragmatic irritation from the surgery as well as subdiaphragmatic surgical drains may explain this symptom, which is difficult to control with epidural management. Removal of all epidural catheters was uncomplicated. No transfusions of blood products to correct hemostasis before catheter removal were necessary. However, it was impossible from our retrospective chart review to determine whether some of our patients may have remained on their epidural infusions longer to reach the necessary INR threshold (INR of ≤ 1.3 per departmental policy) for epidural catheter withdrawal. The current literature suggests that epidural catheters may be removed relatively safely from patients on Coumadin when the INR is less than 1.5.24 Because of the complex postoperative changes in hemostasis, it remains to be determined whether this standard also may be applied safely to this donor population. Close clinical and laboratory monitoring of these patients for neurologic or infectious complications when using catheter-based neuraxial pain control is mandatory. Delaying the removal of epidural catheters until resolution of the postoperative coagulopathy or administering donated blood products for its correction add unwelcome risk, time, and expense to the hospital course but may not always be avoidable.

Limitations of our study are its retrospective design and its small sample. The contribution of the local anesthetic field infiltration to pain control is difficult to determine, and the influence of early right shoulder pain on VAS scores in phase 1 is unclear. Conclusions on the safety of epidural catheters in these patients cannot be generalized. Secondary outcomes such as time to ambulation, return of bowel function, and occurrence of urinary retention also would be of great interest but were not captured. However, our results still may be able to offer some insight into the perioperative issues of these donors, serve as a stimulus for debate, and guide the design of future trials.

Until further definition of hepatectomy-induced physiologic alterations, effective pain control alternatives that do not rely on an epidural catheter are desirable. Multimodal analgesia may offer such an alternative. The term multimodal analgesia or more recently combination analgesic chemotherapy, refers to the simultaneous use of multiple analgesic drugs or methods that maximize analgesic benefit by targeting multiple nociceptive pathways.17, 25 Safe and effective drug dosing, including that of local anesthetic field infiltration or continuous incisional local anesthetic delivery via a catheter based system, should be explored in this donor group.

In conclusion, field infiltration and epidural pain management was effective and safe for postoperative analgesia in this case series. Universally, coagulopathy accompanied right hepatectomy for living liver donation, mandating careful neurologic and laboratory monitoring if neuraxial pain control is employed. Although the potential risk of epidural pain control does not appear to be prohibitive in this population, it is not possible to quantify this risk currently. Balancing the many merits of optimal analgesia versus the risks associated with current techniques to achieve these benefits requires a body of clinical trial evidence that does not yet exist. Although it is debatable whether the risk of epidural analgesia is justifiable in the pain management of these organ donors, alternative options may introduce different safety concerns and may not be as effective. Future research objectives include the following: (1) characterization of hepatic dysfunction and altered pharmacokinetics and dynamics, (2) clarification of the safety of epidural pain management in these patients, (3) elimination of “analgesic gaps” during postoperative pain management, and (4) determination of effective multimodal alternatives to a primarily epidural catheter-based pain control regimen. Achievement of these goals could be hastened by creation of a national living liver donor management and perioperative outcomes database.