Evaluation of standard liver volume formulae for Chinese adults

Abstract

AIM: To evaluate different standard liver volume (SLV) formula and verify the applicability of the formulae for Chinese adults.

METHODS: Data from 70 cases of living donor liver transplantation (LDLT) performed at our transplantation centers between January 2008 and April 2009 were analyzed. SLV was estimated using our recently reported formula [the Chengdu formula: SLV (mL) = 11.5 × body weight (kg) + 334] and other reported formulae used for Chinese adults. Actual intraoperative liver volumes were obtained from a review of the patients’ medical records.

RESULTS: The actual right liver volume was not significantly different from the estimated right liver volume determined by the Chengdu formula, but was significantly smaller than estimates using the Heinemann, Urata, Vauthey, and Lee formulae (P < 0.01), and significantly larger than estimates using the Fan formula (P < 0.05).

CONCLUSION: The Chengdu formula was demonstrated to be reliable by its application in LDLT.

Key Words: Standard liver volume; Living donor liver transplantation; Chinese adult; Liver volume formula

INTRODUCTION

Living donor liver transplantation (LDLT) has been used to alleviate the shortage of available liver donors. Accurate estimation of the standard liver volume (SLV) of the living donor and recipient is crucial. Overestimation of the donor’s SLV may result in excessive hepatic resection leading to liver failure, while underestimation of the recipient’s SLV may result in small-for-size graft syndrome[1–5]. Since 2001, our transplant centers have carried out 212 LDLTs. We estimated the SLV using computed tomography (CT) or reported formulae. However, there was a difference between these estimates and the actual liver volumes (ALVs) for Chinese adults. Recently, we developed a new formula (named the Chengdu formula) to estimate SLV using data from 115 LDLTs[6]. The formula is: SLV (mL) = 11.5 × body weight (kg) + 334. Using this formula, the SLVs were evaluated in 76 cases of LDLT performed from January 2008 to April 2009. Its accuracy was compared to that of other internationally reported formulae[7–10] to assess which formula is the most accurate for Chinese adults.

MATERIALS AND METHODS

Patient selection

The data from 76 living donors were analyzed. Inclusion criteria were: (1) a healthy adult donor, aged 19-59 years; (2) right liver graft without middle hepatic vein; (3) adult-to-adult LDLT; (4) single donor; (5) no history of long term drinking. Exclusion criteria: (1) donor age < 18 or > 60 years; (2) left hepatic graft or left lateral lobe graft; (3) double donor grafts; (4) adult-to-child transplant; (5) donors who were hepatitis B or C carriers[11–14].

Clinical data

Data of preoperative donors included age, sex, height (BH, measured to the nearest 1 cm), body weight (BW, measured to the nearest 0.5 kg), and body surface area (BSA) calculated using the DuBois formula: BSA (m²) = BW (kg) 0.425 × BH (cm) 0.725 × 0.007184 or the Mosteller formula: BSA (m²) = square root BH (cm) × BW (kg)/3600. From the diaphragm to the superior mesenteric artery plane, the entire liver image was scanned using a 7 mm thick layer. In the Leonardo workstation, the LV was measured by venous phase images[1516]. All preoperative CT examinations of donors were performed by a single radiologist and all donor procedures were performed by the same surgical unit. The volume of the grafts was measured by a 3 L beaker using a drainage method intraoperatively and the error was less than 10 mL[1718].

Right liver graft without middle hepatic vein reconstruction from a living donor was performed as described, with temporary occlusion of the right portal vein (PV) and right hepatic artery and use of ultrasonography to guide parenchymal transection. The right hepatic duct, right hepatic artery, right portal vein branch, and right hepatic vein were transected approximately 2-3 mm from the confluence[1920], leaving the donor’s main PV and confluence intact. The graft was flushed with University of Wisconsin solution through the PV and hepatic artery [2122].

The volume of 70 livers was calculated using the Chengdu standard LV formula[6] as described above. The estimated right LV (ERLV) was obtained by multiplying the SLV by the proportion of the LV contributed by the right lobe on CT. The actual right LV (ARLV) was obtained by intraoperative measurement. The difference between the ERLV and ARLV was statistically evaluated.

The formulae of Heinemann et al[8], Urata et al[7], Vauthey et al[9], Lee et al[5], and Fan et al[4] in addition to our own formula[6] were used to determine the estimated SLV (ESLV) of our donor livers. The previously reported formulae are shown in Table 1. For each liver, we calculated the difference between the ALV and volume estimated by each formula (ELV).

Table 1 Reported formulae for ESLV.

Author	Report date	Formula	Material used (race, number)
Urata et al[7]	1995	ESLV = 706.2 × BSA + 2.4	CT Volumetry (Japanese, 96)
Heinemann et al[8]	1999	ESLV = 1072.8 × BSA - 345.7	Autopsy (Caucasian 1332)
Vauthey et al[9]	2002	LV = 18.51 × BW + 191.8	CT volumetry (Western, 292)
Lee et al[5]	2006	ESLV = 691 × BSA + 95	LDLT (Korea, 311)
Fan et al[4]	2000	ESLW = 218.32+BW × 12.29 + gender × 50.74 (M = 1, F = 0)	LDLT (Chinese, 159)
Chengdu[6]	2009	ESLV = 334.024 + 11.508 × BW	LDLT (Chinese, 115)

ESLV: Estimated standard liver volume; BSA: Body surface area; BW: Body weight; CT: Computed tomography; LDLT: Living donor liver transplantation.

Statistical analysis

After testing for normal distribution (kurtosis and skewness tests), descriptive statistics were calculated and data were expressed as means ± SD for age (year), BW (kg), BH (cm), body mass index (BMI), and BSA. The ERLV-ARLV and the ELV-ALV were compared by the 2-sided paired-samples t-test. P < 0.05 was considered statistically significant. All statistical analyses were performed using the SPSS (version 13.0) program.

RESULTS

Seventy donors (all Chinese; 53 men and 17 women; mean age, 32.21 ± 10.07; range, 19-57 years) met the selection criteria. All donors were related to the recipients.

The characteristics of donors are shown in Table 2. All donors were considered healthy on the basis of BMI. All but one donor with a BMI of 17 kg/m² had a BMI of 18-28 kg/m². The mean volume of the right lobe on CT was 658.98 ± 81.14 mL and represented 55.4% ± 3.7% of the whole liver on CT.

Table 2 Donor characteristics.

Age (yr)	32.21 ± 10.07 (19-59)
Gender (Male:Female)	53:17
Body weight (kg)	62.97 ± 8.41 (42-87)
Body height (cm)	167.31 ± 8.15 (148-185)
Body mass index (kg/m2)	22.23 ± 2.44
Body surface area (m2) by DuBois formula	1.7082 ± 0.14
Body surface area (m2) by Mosteller formula	1.7081 ± 0.14
Total liver volume on CT (mL)	1189.53 ± 114.75
Right lobe graft volume on CT without MHV	658.98 ± 81.14
Right lobe volume without MHV to total liver volume on CT (%)	55.4 ± 3.7
Actual right liver volume (mL)	578.58 ± 72.33

MHV: Middle hepatic vein.

The mean ELV and mean ERLV using the Chengdu standard formula were 1058.70 ± 96.74 mL and 586.15 ± 67.17 mL, respectively. The mean ARLV was 578.58 ± 72.33 mL. Differences for individual donors between ERLV and ARLV were not significant (t = -1.882, P = 0.064). A plot of the relationship of ARLV to the ERLV calculated using the Chengdu formula is shown in Figure 1.

Figure 1 Correlation between actual right liver volume (ARLV) and estimated right liver volume (ERLV) by the Chengdu formula. When both were the same, a dot would be on the linear line.

The mean total LV determined preoperatively on CT was 1189.53 ± 114.75 mL. The mean RLV on CT without the middle hepatic vein was 658.98 ± 81.14 mL, and 55.4% ± 3.7% of the total LV. The ALV calculated from the volume of the graft and the ratio of the RLV to the total LV on CT (%) was 1050.10 ± 107.41 mL. The Heinemann, Urata, Vauthey, and Lee formulae significantly overestimated the LV (P < 0.01), while the Fan formula significantly underestimated the LV (P < 0.05). There was no significant difference between ALV and ELV using the Chengdu formula (Figure 2).

Figure 2 Correlation between actual liver volume (ALV) and estimated liver volume (ELV) by each formula. When both were the same, a dot would be on the linear line. Formulae of Urata, Heinemann, Vauthey, and Lee overestimated LV with respect to ALV. The Fan formula underestimated LV and the Chengdu formula gave a good estimate of ALV.

DISCUSSION

CT has become a standard method for assessing liver graft volume in living donors. Estimation of LV by CT (compared to actual volume) has a margin of error of 5%-25%[2324]. In the present study, all donors had preoperative CT assessment of LV (mean total LV, 1189.53 ± 114.75 mL and mean volume of right lobe graft without middle hepatic vein, 658.98 ± 81.14 mL). The actual volume of the right liver was 578.58 ± 72.33 mL. In the present study, the LV on CT was 10%-20% higher than the ALV[25–27]. The reasons may be as follows: (1) Preoperative CT measurement is carried out under normal blood flow conditions. Perioperatively, liver resection interrupts the blood supply causing a loss of liquid volume, collapse of supporting structures, and thereby reduction in the volume of the liver. (2) Sources of error (partial volume effect, inter-observer variation, and respiratory movements) may account for this difference[28].

The difference between the ERLV (using our formula) and ARLV was compared to the difference between ERLV, calculated using the formulae of Heinemann, Urata, Vauthey, Lee, and Fan, and ARLV in our 70 donors. The Heinemann, Urata, Vauthey, and Lee formulae overestimated LV (P < 0.01)[29]. The reasons may include: ethnic differences (patients in Europe and the United States were Caucasian). All except the Sheung Tat Fan and Chengdu formulae were used to estimate LV from CT LV or autopsy LV. Estimates of LV by CT were 5%-25% higher than the ALV[30].

Statistical analysis showed that the Fan formula tends to underestimate LV. The weight and height of the donors in our study were higher than of those in the Hong Kong group. This may be one of the reasons both results are very close (Table 3). Above all, we believe that the Chengdu formula was demonstrated to be reliable by its application in LDLT. We were limited to use of single center data in the present study, but we hope to improve the formula by using national multicenter data in the future[31].

Table 3 Statistical analysis of estimated LV by each formula.

Formula	mean ± SD	t	P-value
Urata	1208.73 ± 99.92	-20.91	P < 0.01
Heinemann	1486.85 ± 151.78	-40.84	P < 0.01
Vauthey	1357.40 ± 155.60	-32.44	P < 0.01
Lee	1275.36 ± 97.77	-29.87	P < 0.01
Fan	1034.28 ± 111.61	2.465	P = 0.016
Chengdu	1058.70 ± 96.74	-1.417	P = 0.161
ALV	1050.10 ± 107.41	ND	ND

ALV: Actual liver volume; ND: Not determined.

COMMENTS

Background

With development of living donor liver transplantation (LDLT), especially improvement of right graft adult-to-adult LDLT, the danger of donating has been paid more and more attention. The exact liver volume is not only relevant for the recipient, but also for the donor to avoid dangerous life-threatening residual liver volumes.

Research frontiers

Scholars of different countries established several standard liver volume (SLV) formulae from clinical data. The authors estimated the SLV using computed tomography or reported formulae. However, there was a gap between these estimates and the actual liver volumes for Chinese adults. Recently, they developed a new formula (named the Chengdu formula) to estimate SLV using data from 115 LDLTs.

Innovations and breakthroughs

With the Chengdu formula, the SLVs were evaluated in 76 cases of LDLT performed from January 2008 to April 2009. Its accuracy was compared to that of other internationally reported formulae to assess which formula is the most accurate for Chinese adults.

Applications

With national multicenter data in the future, the Chengdu formula for SLV can be improved. It may then be applied to the evaluation of donors for LDLT.

Terminology

Standard liver volume: normal liver volume without disease affecting the volume of liver.

Peer review

Very interesting manuscript dealing with a very hot topic: determination of optimal size matching between graft and recipient in LDLT by means of race-adapted calculation of liver volumes. The recently published liver volume formula for Chinese people (Chengdu formula) has been demonstrated to be more reliable than others and therefore it should be adopted especially in this particular form of LT.