Microvascular transfer of free musculocutaneous flaps has become a routine procedure for reconstruction of the lower extremity after major trauma or extensive tumour surgery [1-2]. Despite improved surgical techniques, partial or total loss of the flaps remains an important complication [3-5]. Although surgical skill is probably the most important factor determining the outcome of microvascular surgery other factors such as anaesthetic management also play an important role . General anaesthesia can in many ways influence local and regional blood flow in different organs and tissues including free flaps [6-8]. Because free flap surgery is a long procedure, general anaesthesia is usually recommended for the patients' comfort.
It has been suggested that regional anaesthesia improves blood flow in free flaps [9-12] and that it may prevent vasospasm at the microvascular anastomoses . On the other hand, there have been concerns that vasodilatation produced by regional anaesthesia might be more pronounced in normal innervated tissues than in the tissues of the free flap, which could result in decreased blood flow in the flap: a steal phenomenon. In two cases reported in the literature, epidural block caused a decrease in free flap blood flow [6,14]. In addition, in a recent experimental study in pigs, in which clinical conditions for anaesthesia and microvascular surgery were imitated as closely as possible, it was found that blood flow in free musculocutaneous flaps to the lower extremity was significantly decreased after epidural anaesthesia .
Thus, we measured the effect of epidural anaesthesia on blood flow in free flaps to the lower extremity in patients undergoing reconstructive microvascular surgery under general anaesthesia. We compared changes in microcirculatory blood flow in skin and/or muscle of free flaps with changes in blood flow of intact skin and muscle on the same extremity by the use of a multichannel laser Doppler flowmetry [3,16,17]. Another group of patients undergoing free flap surgery without epidural anaesthesia was studied to detect any spontaneous changes in the systemic blood pressure and microcirculatory blood flow (in free flap-and control tissues) during the time frame of the study, i.e. up to 40 min after completion of surgery.
Twenty-one consecutive adult patients (ASA I and II), scheduled for elective free flap transfer to the lower extremity and for post-operative pain therapy with epidural anaesthesia were included in the study. Another seven patients undergoing the same type of surgery without epidural anaesthesia were studied to monitor whether or not the systemic and local haemodynamics remained constant during the study period. Informed consent was obtained from all the patients and the protocol was approved by the local ethics committee.
Patients' data are shown in Table 1. In the epidural group, eight patients were scheduled for free gracilis muscle flaps, six for free fasciocutaneous flaps and seven for free musculocutaneous flaps. The seven patients who did not receive epidural anaesthesia were scheduled for free musculocutaneous flaps.
Before induction of general anaesthesia, a lumbar epidural catheter (L2-L3 or L3-L4) was inserted for routine post-operative pain therapy in the test group. The initial dose of 10-15 mL of lignocaine hydrochloride 2% was administered by fractional injection and adjusted to body weight and patient size. The spread and efficacy of the epidural block were verified by a pin-prick test 10, 15 and 20 min after administration of the local anaesthetic. Desired spread of the epidural block was considered at least one spinal segment above and one below the planned recipient site for the free flap. No further drugs were administered through the catheter until after completion of surgery. The control group did neither receive an epidural catheter nor epidural anaesthesia. Body temperature of all the patients was maintained above 36°C during surgery by raising the operating room temperature to 26°C and by using a forced warm air system with appropriate blankets (BairHugger®, Augustine Medical, Eden Prairie, MN, USA).
After completion of surgery, which was performed under general anaesthesia (isoflurane, N2O-O2, fentanyl, pancuronium) and lasted 4-8 h (mean 328 min), light general anaesthesia with isoflurane (0.5±0.1% end-tidal concentrations) in nitrous oxide-oxygen (70-30) was continued during the haemodynamic measurements on all patients. All the patients were covered with warming blankets during the study period and end-tidal CO2 and isoflurane concentrations were kept constant until all measurements were completed. Following a stabilization period of at least 15 min after surgery, base-line measurements (t=−15 min) were performed on all patients after which the patients with epidural anaesthesia received a rapid intravenous (i.v.) infusion of 10 mL kg−1 Ringer's lactate during the 15 min starting immediately after base-line measurements (15 min before induction of the epidural block), while the patients who did not receive epidural anaesthesia received 2 mL kg−1 Ringer's lactate during the 15 min after base-line recordings. Fifteen minutes after base-line recordings (t=0 min), the dose of lignocaine hydrochloride 2% (10-15 mL) determined preoperatively was injected through the lumbar epidural catheter in the test group. The local anaesthetic was administered in two steps; a test dose (20% of the total amount) was followed 3 min later by the remaining quantity. Arterial hypotension (MAP ≤60 mmHg) was immediately treated with an i.v. infusion of colloids (6% hydroxyethyl starch 200/0.5) up to 15 mL kg−1. In addition, ephedrine 10 mg was given i.v. if MAP remained below 60 mmHg after i.v. fluid administration. From t=0 min until t=20 min microcirculatory blood flow and arterial blood pressure were measured continuously. Pilot experiments performed before designing the present study showed that the maximum changes in microcirculatory blood flow and in arterial blood pressure after the epidural administration of lignocaine hydrochloride 2% occurred after 10-15 min and these changes remained virtually unchanged for at least 60 min. After completion of the measurements, general anaesthesia was discontinued and the patients were allowed to wake up.
Microcirculatory blood flow was measured continuously by the use of two dual channel laser Doppler flowmeters (Periflux 4001 Master, Perimed, Järfälla, Sweden). The time constant of the laser Doppler flowmeter (LDF) output amplifier was set at 3 s and the sampling rate at 10 Hz [18,19]. Angled LDF probes (PF 404) were sutured onto the muscle and/or skin in the central portion of the free flap and two other probes were sutured to the surrounding intact skin and muscle (control skin and control muscle, respectively) on the same extremity.
Arterial blood pressure (MAP, mmHg) and heart rate (HR, bpm) were recorded from a radial artery catheter and ECG, respectively. End-tidal CO2 (ET-CO2, %) and end-tidal isoflurane (ET-Iso, vol%) concentrations in addition to body temperature (rectal probe) were measured continuously using a multimodular vital sign monitor (Hellige-Marquet, Freiburg, Germany).
Data acquisition and analysis
Microcirculatory blood flow and blood pressure data were acquired on-line via a multichannel interface (Mac Paq MP 100; Biopac Systems, Goleta, CA, USA) with acquisition and analysis software (Acqknowledge 3.0™; Biopac Systems, Goleta, CA, USA) to a portable computer (PowerBook, Apple Computer, Cupertino, CA, USA). Microcirculatory blood flow measured with LDF is presented as percentage of the base-line values. Data for microcirculatory flow and MAP before and after induction of the epidural block represent the average of 5 min immediately before base-line (t=−15 min) and average of 5 min 10-15 min after injection of lignocaine via the epidural catheter. The data are presented as mean±SEM. P < 0.05 was considered statistically significant. The InStat 2.03® statistical package (GraphPad Software, CA, USA) was used for statistical analysis.
The study was designed to compare changes in the microcirculatory blood flow in skin and muscle of free flaps with changes in intact skin and muscle of the same extremity during induction of epidural anaesthesia. The intact skin and muscle in the epidural group acted as control tissue. The purpose of the group not receiving epidural anaesthesia was merely to establish whether or not systemic arterial blood pressure and microcirculatory flow remained stable during the time frame of the study i.e. until 40 min post-operatively when no epidural block was induced. Therefore, no attempt was made to perform statistical comparisons between the epidural group and the group which did not receive epidural anaesthesia. The statistical analysis was performed between base-line values and values at 15 min in flap tissues as well as between base-line and values at 15 min in intact tissues using the two-tailed paired Student's t-test. In addition, comparison was made between changes in flow in flap tissues and flow in intact tissue (control skin and muscle) in the same patients using the two-tailed unpaired Student's t-test. The correlation coefficient r was calculated by using the parametric Pearson correlation test.
The epidural block caused a marked decrease in microcirculatory blood flow in all the flaps studied (Fig. 1). In muscle flaps, it was reduced to 80±8% (P < 0.05) and in fasciocutaneous flaps to 69±6% (P < 0.05) when compared with base-line values. In musculocutaneous flaps flow was reduced to 79±5% (P < 0.05) in muscle and to 74±4% (P < 0.05) in skin. In intact control muscle and skin of the same extremity microcirculatory flow remained virtually unchanged despite a significant decrease in blood pressure (see below).
After induction of the epidural block, MAP decreased from 85.0±2.9 mmHg to 68.5±2.7 mmHg (P < 0.01), while HR, end-tidal CO2, and isoflurane concentrations remained constant (Table 2). There was a significant positive correlation between changes in microcirculatory flow in the flap and changes in MAP occurring after induction of the epidural block (y=0.95x+1.29 for microcirculatory flow as a function of MAP; P < 0.01; r=0.71). The lowest blood pressure was observed 9-13 min after injecting the local anaesthetic. Eight patients required additional i.v. fluids to keep MAP over 60 mmHg and three of these also received an i.v. injection of ephedrine. Administration of ephedrine caused a marked increase in microcirculatory flow (19%; 70%; 145%) in the free flap and in MAP (23%; 30%; 56%).
In the patients who did not receive epidural anaesthesia, no significant changes were recorded in MAP, microcirculatory flow in the free flaps or in the control tissues during the equivalent time period (40 min; Table 2 and 3).
Mean duration of surgery was 328±16 min, mean intra-operative blood loss 698±16 mL and haematocrit at the end of surgery was 33±1.1%. Spread of the epidural block was recorded in the recovery room (using pin prick) within 30-40 min of finishing the study. It was found that the epidural block covered the area of the free flap and the control skin and muscle in all patients by a margin of at least two spinal segments above and one below. The highest level recorded was T5 in one patient.
Despite immediate surgical revision, one latissimus dorsi flap necrotised due to thrombosis of the venous anastomoses 56 h after surgery. Partial necrosis occurred in the skin island of the musculocutaneous rectus femoris flap and the distal portion of one gracilis muscle flap.
The results of this clinical study do not support the hypothesis that epidural anaesthesia improves microcirculatory blood flow in free flaps. On the contrary, induction of epidural block caused a 20-30% decrease in microcirculatory flow in all three types of free flaps investigated, while flow in intact control tissues (muscle and skin) on the same extremity remained unchanged (Fig. 1). Considering that raising a musculocutaneous flap alone decreases its blood flow by approximately 20%  an additional decrease in blood flow caused by epidural anaesthesia may become critical to survival of the flap. The present results are in agreement with those recorded in a recent experimental study using pigs in our laboratory in which clinical conditions for anaesthesia and microvascular surgery were imitated as far as possible . It was shown that epidural anaesthesia decreased total blood flow (regional flow) significantly as well as microcirculatory blood flow in free musculocutaneous flaps despite the fact that cardiac output and tissue blood flow in control muscle and skin remained constant.
One can speculate why epidural anaesthesia did not produce the 'desired' effects on blood flow in free flaps. The fall in systemic arterial blood pressure from 85.0 mmHg to 68.5 mmHg may appear as a likely cause. However, mean arterial blood pressure of 68.5 mmHg in ASA class I and II patients should be more than sufficient to maintain normal perfusion of all body tissues except in the presence of some relative or absolute hypovolaemia. The fact that the blood pressure and blood volume were sufficient to maintain unchanged microcirculatory blood flow in control skin and muscle does not suggest hypovolaemia. Furthermore, in the experimental study cited above  induction of epidural anaesthesia also resulted in a fall in blood pressure, which was accompanied by a decreased flow in free musculocutaneous flaps though the cardiac output and control muscle and skin blood flow remained constant.
The feeding artery to the free flaps at the recipient site has an intact sympathetic innervation and it is therefore influenced by epidural anaesthesia. However, arteries of this size (1-3 mm in diameter) are known to have negligible effects on vascular resistance and blood flow. Vascular resistance is controlled by sympathetic tone at the arteriolar level [20,21]. The sympathetic block caused by epidural anaesthesia can be expected to have an effect only on the arterioles in intact skin and muscle because the tissues in the free flaps are denervated. Consequently, it may be assumed that the epidural block resulted in reduced vascular resistance in the intact control tissues of the lower extremity while it remained unchanged in the denervated flap tissues. This may have caused redistribution of blood flow from the free flap to the surrounding tissues, which were vasodilated by the epidural block: a steal phenomenon [6,8,14,15]. This assumption is partly supported by a recent experimental study in pigs in which an i.v. infusion of sodium nitroprusside caused a marked decrease in microcirculatory blood flow in free musculocutaneous flaps but the blood flow in control tissues and the cardiac output remained unchanged .
There were two important reasons for studying the effects of epidural block under general anaesthesia and not in a wakening state post-operatively. Firstly, it has been suggested that it could be beneficial to use general anaesthesia combined with epidural anaesthesia for microvascular surgery and we hoped to clarify this hypothesis [6,9-12]. Secondly, in awake patients it may be difficult to control many factors that can influence both the macro- and the microcirculation during the study period, such as nausea, vomiting, anxiety and pain. Furthermore, induction of epidural anaesthesia frequently causes shivering in awake patients, which can result in sudden changes in cardiac output, arterial blood pressure and microcirculatory flow. Such events would make interpretation of the results more difficult. Under general anaesthesia, on the other hand, it was possible to avoid these undesirable effects. Although isoflurane, as in the case of other volatile anaesthetic drugs, may reduce sympathetic tone and alter local/regional blood flow, the doses of isoflurane used in the present study were low and should therefore have had only minimal haemodynamic effects [6,23,24]. It has also been shown recently that both total blood flow and microcirculatory blood flow are well maintained in musculocutaneous flaps in an experimental study in pigs under isoflurane anaesthesia, even in the presence of hypovolaemia . Furthermore, possible effects of general anaesthetics cannot explain the difference in flow between free flaps and intact control tissue in the same patients.
It was our aim to measure the effects of a 'complete' epidural block on microcirculatory blood flow in free flaps. Therefore, we used a 'full strength' local anaesthetic (2% lignocaine) to cause a complete sensory and sympathetic blockade. It is likely that a less concentrated solution of local anaesthetic, as used frequently for post-operative pain, or opiates given epidurally would have caused different effects on flap blood flow [25,26]. Therefore, epidural use of opiates alone or combined with diluted local anaesthetics may be more suitable for patients undergoing microvascular surgery if the main purpose of epidural drug administration is post-operative pain therapy  and not the interruption of painful stimuli during surgery.
It is concluded that epidural anaesthesia combined with general anaesthesia may significantly reduce microcirculatory blood flow in free muscle and musculocutaneous flaps.
This study was supported by the Swiss National Foundation for Scientific Research (grant #0032-040761) and by a grant from the Sandoz Research Foundation, Basel.
1 Khouri RK, Shaw WW. Reconstruction of the lower extremity with microvascular free flaps
. J Trauma
2 Khouri RK. Free flap surgery: The second decade. Clin Plastic Surg
3 Hovius S, van Adrichem L, Mulder H, van Strik R, van der Meulen J. The predictive value of the laser Doppler flowmeter for postoperative microvascular monitoring. Ann Plastic Surg
4 Hidalgo DA, Jones CS. The role of emergent exploration in free tissue transfer. A review of 150 consecutive cases. Plastic Reconstr Surg
5 Khouri RK. Avoiding free flap failure. Clin Plastic Surg
6 Sigurdsson GH, Thomson D. Anaesthesia and microvascular surgery
: clinical practice and research. Eur J Anaesth
7 Pang CY, Forrest CY, Morris SF. Pharmacological augmentation of skin flap viability: a hypothesis to mimic the surgical delay phenomenon or a wishful thought. Ann Plastic Surg
8 Sigurdsson GH, Banic A, Wheatley AM, Mettler D. Effects of halothane and isoflurane anaesthesia on microcirculatory blood flow in musculocutaneous
flaps. Br J Anaesth
9 Robins DW. The anaesthetic management of patients undergoing free flap transfer. Br J Plastic Surg
10 Scott GR, Rothkopf DM, Walton RL. Efficacy of epidural anesthesia in free flaps
to the lower extremity. Plastic Reconstr Surg
11 Strecker WB, Wood MB, Wedel DJ. Epidural anesthesia during lower extremity free tissue transfer. J Reconstr Microsurg
12 Stanisich V, Butel HC. Anaesthesia in microvascular surgery
. In: OBrien BM, Morrison WA, eds. Reconstructive Microsurgery.
Edinburgh: Churchill Livingstone, 1987: 129-133.
13 Phelps DB, Rutherford RB, Boswick JA. Control of vasospasm following trauma and microvascular surgery
. J Hand Surg
14 van Twisk R, Gielen WM, Pavlov PW, Robinson PH. Is additional epidural sympathetic block in microvascular surgery
contraindicated? A preliminary report. Br J Plastic Surg
15 Banic A, Krejci V, Erni D, Petersen-Felix S, Sigurdsson GH. Effects of extradural anesthesia on microcirculatory blood flow in free latissimus dorsi musculocutaneous
flap in pigs. Plastic Reconstr Surg
16 Banic A, Sigurdsson G, Wheatley A. Continuous perioperative monitoring of microcirculatory blood flow in pectoralis musculo-cutaneous flaps. Microsurg
17 Clinton M, Sepka R, Bristol D et al.
Establishment of normal ranges of laser Doppler blood flow in autologous tissue transplants. Plastic Reconstr Surgery
18 Kuznetsova L, Tomasek N, Sigurdsson GH, Erni D, Banic A, Wheatley A. Estimation of muscle blood flow and resistance by laser Doppler flowmetry
during changes in vascular tone. Am J Physiol
19 Löfström JB, Cousins MJ. Sympathetic neural blockade of upper and lower extremity. In: Cousins MJ, Bridenbaugh PO, eds. Neural Blockade in Clinical Anesthesia and Management of Pain.
Philadelphia: J. B. Lippincott, 1988: 461-500.
20 Roddie JC. Circulation to skin and adipose tissue. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology.
Bethesda: American Physiology Society, 1983: 285-318.
21 Shepherd JT. Circulation to skeletal muscle
. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology.
Bethesda: American Physiology Society, 1983: 319-370.
22 Banic A, Krejci V, Erni D, Wheatley A, Sigurdsson G. Effects of sodium nitroprusside and phenylephrine on blood flow in free musculocutancous flaps during general anesthesia. Anesthesiology
23 Seyde WC, Longnecker DE. Anesthetic influence on regional hemodynamics in normal and hemorrhaged rats. Anesthesiology
24 Weiskopf RB, Townsley MI, Riordan KK, Chadivick K, Baysinger M, Mahoney E. Comparison of cardiopulmonary responses to graded hemorrhage during enflurane, halothane, isoflurane and ketamine anesthesia. Anesth Analg
25 Breivik H. Safe perioperative spinal and epidural analgesia: importance of drug combinations, segmental site of injection, training and monitoring. Acta Anaesthesiol Scand
26 Cousins MJ, Bromage PR. Epidural neural blockade. In: Cousins MJ, Bridenbaugh PO, eds. Neural Blockade in Clinical Anestliesia and Management of Pain.
Philadelphia: J. B. Lippincott, 1988: 253-360.