Profile of a randomized controlled trial.
Hasegawa K, Takayama T, Orii R, Sano K, Sugawara Y, Imamura H, Kubota K, Makuuchi M. Effect of Hypoventilation on Bleeding During Hepatic ResectionA Randomized Controlled Trial. Arch Surg. 2002;137(3):311-315. doi:10.1001/archsurg.137.3.311
Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
Blood loss in hepatic resection is an important determinant of operative outcome.
To clarify whether reducing the tidal volume would be effective in decreasing blood loss during liver transection.
Randomized controlled trial.
Eighty patients scheduled to undergo hepatic resection were randomly assigned to receive liver transection under normoventilation (n = 40) or hypoventilation (n = 40).
During liver transection, in the normoventilation group, the tidal volume was 10 mL/kg and the respiratory rate was 10/min; in the hypoventilation group, the tidal volume was reduced to 4 mL/kg and respiratory rate was increased to 15/min. Liver transection was performed under total or selective inflow occlusion.
Main Outcome Measure
Between the normoventilation and hypoventilation groups, no significant difference was found in total blood loss (median [range]: 630 mL [72-3600 mL] vs 630 mL [120-3520 mL]; P = .44) or blood loss per transection area (median [range]: 7.3 mL/cm2 [1.2-55.4 mL/cm2] vs 9.8 mL/cm2 [0.9-79.9 mL/cm2]; P = .55). During liver transection, the central venous pressure was significantly reduced in the hypoventilation group than in the normoventilation group (median [range]: –0.7 cm H2O [–3.0 to 1.8 cm H2O] vs –0.2 cm H2O [–4.0 to 2.0 cm H2O]; P = .007). The maximum end-tidal carbon dioxide level in the hypoventilation group was significantly higher than that in the normoventilation group (maximum [range]: 50 mm Hg [28-66 mm Hg] vs 37 mm Hg [27-60 mm Hg]; P<.001). Transection time, postoperative liver function, hospitalization length, morbidity, and mortality were similar in the 2 groups.
This randomized trial suggested no beneficial effect of reduction of tidal volume on bleeding during hepatic resection.
THE SUCCESS of hepatic resection is greatly influenced by the amount of intraoperative hemorrhage, because massive blood loss is associated with poor operative and long-term outcomes.1- 3 One of the greatest interests of liver surgeons is how to reduce blood loss during liver transection. The use of inflow occlusion techniques has made an important contribution to the reduction of blood loss,4,5
but these techniques cannot control hemorrhage from the hepatic veins and their branches. Occlusion of both outflow and inflow, for example, by total vascular exclusion6 and clamping of the inferior vena cava or the hepatic veins,7,8 has been developed. However, the indications for these techniques are limited because they have a profound influence on systemic hemodynamics and involve complicated procedures.9
Clinical data have suggested that central venous pressure (CVP) is positively correlated with the amount of hemorrhage that occurs during liver transection.10,11 Keeping a low CVP, obtained by administration of vasodilators12,13 or limitation of infusion,14,15 has been reported to be effective in minimizing blood loss, but with such methods it is neither reproducible nor easy to control CVP. We hypothesized that reducing the respiratory tidal volume could result in reduced blood loss from the hepatic veins during liver transection. Under positive pressure ventilation, lung inflation directly compresses the heart,16 which is fixed in the mediastinum. Animal experiments have confirmed that pressure in the right atrium is increased by lung inflation,17 and the CVP, which is nearly equal to the pressure in the right atrium, is also increased by lung inflation.18 Reduction of the tidal volume might therefore provide a low CVP by decreasing the pulmonary compression effect. We conducted this randomized controlled trial to clarify whether reducing the tidal volume would be effective in decreasing blood loss during liver transection.
Patients scheduled to undergo hepatic resection for the removal of tumors were entered into this trial. Patients with severe pulmonary dysfunction (<70% vital capacity, or 1 second forced expiratory volume divided by forced vital capacity equaling less than 60%) were excluded. We obtained informed consent from the patients and approval from the local ethics committee.
In the operating room, eligible patients were randomly assigned to the normoventilation or hypoventilation groups by the minimization method, with stratified factors of age (<60 or ≥60 years), type of surgical procedure (minor or major hepatectomy), and indocyanine green clearance at 15 minutes (<20% or ≥20%). Surgical procedures were selected according to the serum total bilirubin level and the indocyanine green clearance, as described previously.15 Limited resection or segmentectomy was considered minor hepatectomy, and sectoriectomy or hemihepatectomy was considered major hepatectomy.
The hepatic parenchyma was divided by clamp crushing or ultrasonic dissection.19 One consultant (M.M.) and 5 trainees (T.T., K.S., Y.S., H.I., and K.K.) performed all of the surgical procedures. Inflow occlusion was obtained intermittently by the Pringle maneuver, with 15 minutes of clamping followed by 5 minutes of unclamping, or by a hemihepatic vascular occlusion method with 30 minutes of clamping and 5 minutes of unclamping.4 During surgery in the normoventilation group, the tidal volume was maintained at 10 mL/kg and the respiratory rate was kept at 10/min.20
In the hypoventilation group, the tidal volume was reduced to 4 mL/kg and the respiratory rate was 15/min during inflow occlusion. In clamp-free periods, the respiratory conditions in the hypoventilation group were the same as those in normoventilation group. The positive end-expiratory pressure was kept at 4 mm Hg during the surgery in both groups.
Vecuronium bromide was administered intermittently (0.8 mg/kg hourly). The degree of neuromuscular block was monitored by the response pattern evoked by train-of-four electrical stimulation (50 mA, 2 Hz). Intraoperative infusion was restricted to 4 to 4.5 mL/kg per hour.15
Unless remarkable changes of hemodynamic state emerged, such as hypotension or tachycardia, the infusion rate was not changed, regardless of CVP value. No vasodilators were administered except in cases with hypertension. The blood loss and the transection area of the specimen were measured. We measured the CVP (from a zero point at the midatrial level via a catheter inserted into the superior vena cava) and the end-tidal carbon dioxide (CO2) level before liver transection (preclamp value) every 5 minutes in clamping (in-clamp values) and 30 minutes after liver transection (postclamp value). Difference between mean value at the start of clamping for all intervals and that at the release was calculated for CVP and end-tidal CO2. If the end-tidal CO2 level rose to 60 mm Hg, the anesthesiologist was permitted to alter the assigned respiratory conditions. Only 2 investigators (K.H. and R.O.), who were not involved in the hepatic resections, had seen the results of the randomization procedure, and they were able to decide to alter the respiratory conditions without consulting with the surgeons. All patients were managed in the same way postoperatively.15
The primary outcome measure was blood loss during hepatic resection. The amount of blood loss was measured in total and during liver transection, and the blood loss per transection area was calculated. Secondary measures included CVP, end-tidal CO2 levels, and peak inspiratory airway pressure during clamping, transection time, fluid infusion, postoperative liver function, hospitalization length, and postoperative morbidity and mortality.
At our institute, the mean (SD) blood loss during liver transection has been 400 mL (320 mL).19,21 We hypothesized that reducing the tidal volume could decrease blood loss by 200 mL, a clinically valuable reduction. Forty patients in each group would be required to detect a significant difference with a 2-tailed type 1 error of 5% and a statistical power of 80%.
All analyses were performed on an intention-to-treat basis. The background characteristics and surgical outcome data of the 2 groups were compared by the χ2 test for categorical data and the Mann-Whitney U test for continuous or ordinal data. A logistic regression test was used to estimate the influence of tidal volume on total blood loss (divided by clinically relevant cutoff points: blood loss ≤500 vs >500 mL),19 controlling for 10 potential confounders (Child-Pugh class, indocyanine green clearance, background liver, surgeon, inflow occlusion, thoracotomy, hepatectomy procedure, number of resections, transection area, and CVP). The results were expressed as adjusted odds ratios (ORs) with 95% confidence intervals (CIs) and P values from the likelihood ratio test. Significance was defined as P<.05. Calculations were done with the help of Statview 5.0 computer software (Abacus Concepts Inc, Berkeley, Calif).
From July 14, 1999, through March 10, 2000, 80 patients were randomized; 40 were assigned to the normoventilation group and 40 to the hypoventilation group (Figure 1). One patient in the normoventilation group with a metastatic liver tumor had multiple lymph nodal metastases; he did not undergo hepatic resection and was excluded. Between the 2 groups, there was no significant difference in any background characteristics (Table 1).
The following outcomes are presented as median (range). The total blood loss was the same in the normoventilation and hypoventilation groups (630 mL [72-3600 mL] vs 630 mL [120-3520 mL]; P = .44; Table 2). Whole blood was transfused into 4 patients in the normoventilation group (520, 700, 880, and 880 mL) and 3 patients in the hypoventilation group (440, 660, and 960 mL). Liver transection time was 63 minutes (9-169 minutes) in the normoventilation group and 60 minutes (13-157 minutes) in the hypoventilation group (P = .62). The total amount and rate of fluid infusion were similar between the 2 groups.
The average (minimum and maximum) CVP levels in clamping were similar between the 2 groups (Table 2). However, the difference of CVP in clamping in the hypoventilation group was significantly larger than that in the normoventilation group (–0.7 cm H2O [–3.0 to 1.8 cm H2O] vs –0.2 cm H2O [–4.0 to 2.0 cm H2O]; P = .007). The median (range) peak inspiratory airway pressure during clamping in the hypoventilation group was significantly lower than the normoventilation group (11 mm Hg [6-17 mm Hg] vs 15 mm Hg [10-20 mm Hg]; P<.001).
The maximum (range) end-tidal CO2 levels in the hypoventilation group were significantly higher than those in the normoventilation group (50 mm Hg [28-66 mm Hg] vs 37 mm Hg [27-60 mm Hg]; P<.001). The increase of end-tidal CO2 level in clamping was significantly larger in the hypoventilation group than that in the normoventilation group (7.5 mm Hg [–3 to 22 mm Hg] vs 0.1 mm Hg [–9 to 9 mm Hg]; P<.001). In 7 patients in the hypoventilation group, the tidal volume was increased by an average of 46% (range, 20%-56%) of the arranged volume according to the predefined instructions (Figure 1). Neither air embolism nor remarkable changes in systemic hemodynamics developed during liver transection in both groups.
Postoperative liver function and hospitalization period did not differ between the 2 groups. Postoperative complications developed in 13 patients in the normoventilation group and 16 patients in the hypoventilation group (P = .54), and there were no operative deaths in either group (Table 3).
Tidal volume had no significant relevance to total blood loss (normoventilation vs hypoventilation; adjusted OR, 0.48; 95% CI, 0.15-1.5; P = .20). In addition, blood loss was significantly influenced only by thoracotomy (yes vs no; adjusted OR, 4.08; 95% CI, 1.24-13.5; P = .02); it was not influenced by the other parameters, including CVP (adjusted OR, 1.07; 95% CI, 0.87-1.31; P = .52).
This trial failed to show any significant difference in blood loss between the normoventilation and hypoventilation groups, although the CVP during liver transection was reduced significantly in the hypoventilation group. Reducing the tidal volume significantly increased the end-tidal CO2 but caused no additional risks to the patients. The results of other secondary outcome measures were similar in the 2 groups. We therefore could not clarify any beneficial effect of reducing the tidal volume in hepatic resection.
Under inflow occlusion, hemorrhage during liver transection is, in theory, derived from hepatic venous tributaries.10 Reduction of hepatic venous pressure might be effective for decreasing hemorrhage. Because the hepatic vein is one of the central branches of the inferior vena cava, hepatic venous pressure is positively correlated with CVP. Although other factors such as liver consistency and hepatectomy procedure may affect the marginal hepatic venous pressure, it would be reasonable to assume that a decrease in CVP would contribute to a reduction in blood loss in liver transection through the reduction of hepatic venous pressure. In fact, previous retrospective studies have revealed that CVP is related to blood loss.10,11 We therefore hypothesized that reduction of respiratory tidal volume might be a simple and useful way of obtaining a low CVP by reducing compression of the heart due to lung inflation.
In this trial, the reduction in CVP in the hypoventilation group during clamping was significantly larger than that in the normoventilation group. This indicates that reduction of tidal volume contributed to a decrease in CVP, thus supporting part of our hypothesis. In both groups we tried to achieve low CVP levels, which almost met the proposed criteria that the CVP should be kept at less than 5 cm H2O.10- 14 To reduce the CVP further from such a value would be difficult; this would be one possible reason why the decrease in CVP induced by a reduced tidal volume was small and the CVP levels during clamping were similar in the 2 groups. The similarity in the CVPs of both groups could be the most likely reason for the lack of a significant difference in blood loss. Although a greater reduction in tidal volume might be effective for obtaining a lower CVP, the wisdom of reducing the tidal volume further could be questionable because of the potential risk for patients.
The risk of increased end-tidal CO2 levels and air embolism may be increased by reducing the tidal volume. No air embolism occurred in the 2 groups. A high end-tidal CO2 level (beyond 60 mm Hg) developed in 7 patients of the hypoventilation group; however, in both groups, there were no signs of hyperkinetic circulation (tachycardia or early hypertension) induced by hypercapnia.22 Peak inspiratory airway pressure was decreased by hypoventilation as expected; however, values of the pressure were acceptable in the 2 groups. In addition, the fluid infusion rate, which roughly reflected fluid infusion requirement, was similar in the 2 groups. Although hypoventilation may decrease CVP, the intervention produced little change of hemodynamic state, compared with normoventilation. The lack of complications associated with the hypoventilation group is an indication that reduction of tidal volume does little harm clinically.
In this trial, multivariate analyses showed that thoracotomy was the only independent factor related to increased blood loss, although it would be reasonable to think that thoracotomy decreases CVP and blood loss because the compression effect of right lung inflation is lower. However, animal experiments have demonstrated that the influence of lung inflation on the right atrium in the open chest is the same as in the closed chest,11,12 and our results support these data. We usually perform thoracotomy in patients with tumors located in the cranial part of the liver, such as in segments 7 or 8, or with large tumors that require major hepatectomy. In these cases, hepatectomy itself is usually difficult and the hepatic venous wall is basically exposed lengthwise on the raw surface of the liver. This might be the cause of the increased blood loss, but assessment of its true role needs to be evaluated in other randomized trials.
Total amount of blood loss was low (median, 630 mL) in the 2 groups, which was nearly equal to our previous results (median, 687.5 mL; range, 40-4072 mL).21 Our principles of intraoperative management for reducing blood loss were restricted infusion and sufficient administration of muscle relaxants to keep low CVP.15,21 The surgical skill of our team, added to the management, might have contributed to the good operative results; however, further investigations are needed to clarify that.
Practically, many factors other than hypoventilation will contribute to CVP during hepatic resection, and therefore it would be difficult to identify what factor is most crucial. However, the aim of this trial was not to clarify the issue, but only to examine effectiveness of hypoventilation on blood loss. Because, in theory, factors other than respiratory tidal volume would affect all outcome measures equally in the both groups, the results obtained by this randomized trial can be of clinical value.
We had presumed that reduction of respiratory tidal volume would be a simple and useful way to reduce blood loss during liver transection, but we failed to demonstrate any beneficial effects in this clinical setting.
This work was supported by a grant-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan (grant 12470252) (Drs Kubota and Makuuchi).
Corresponding author and reprints: Masatoshi Makuuchi, MD, PhD, Division of Hepato-Biliary-Pancreatic and Transplantation Surgery, Department of Surgery, Faculty of Medicine, University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8655, Japan (e-mail: email@example.com).