Factors Associated With Postreperfusion Syndrome in Living Donor Liver Transplantation: A Retrospective Study

Abstract

BACKGROUND: 

Postreperfusion syndrome (PRS) after portal vein reperfusion during liver transplantation (LT) has been reported to cause rapid hemodynamic changes and is associated with a prolonged postoperative hospital stay, renal failure, and increased mortality. Although there are some reports on risk factors for PRS in brain-dead donor LT, there are a few reports on those in living donor LT. Therefore, we retrospectively reviewed the factors associated with PRS to contribute to the anesthetic management so as to reduce PRS during living donor LT.

METHODS: 

After approval by the ethics committee of our institution, 250 patients aged ≥20 years who underwent living donor LT at our institution between January 2013 and September 2018 were included in the study. A decrease in mean arterial pressure of ≥30% within 5 minutes after portal vein reperfusion was defined as PRS, and estimates and odds ratio (OR) for PRS were calculated using logistic regression. The backward method was used for variable selection in the multivariable analysis.

RESULTS: 

Serum calcium ion concentration before reperfusion (per 0.1 mmol/L increase; OR, 0.74; 95% confidence interval (CI), 0.60–0.95; P < .001), preoperative echocardiographic left ventricular end-diastolic diameter (per 1-mm increase: OR, 0.90; 95% CI, 0.85–0.95; P < .001, men [versus women: OR, 2.45; 95% CI, 1.26–4.75; P = .008]), mean pulmonary artery pressure before reperfusion (restricted cubic spline, P = .003), anhepatic period (restricted cubic spline, P = .02), and graft volume to standard liver volume ratio (restricted cubic spline, P = .03) were significantly associated with PRS.

CONCLUSIONS: 

In living donor LT, male sex and presence of small left ventricular end-diastolic diameter, large graft volume, and long anhepatic period are associated with PRS, and a high calcium ion concentration and low pulmonary artery pressure before reperfusion are negatively associated with PRS.

KEY POINTS

• Question: What are the factors associated with postreperfusion syndrome (PRS) after portal vein reperfusion in living donor liver transplantation (LDLT)?
• Findings: Male sex and presence of small left ventricular end-diastolic diameter (LVEDD), large graft volume, low calcium ion concentration and high pulmonary artery pressure before reperfusion, and long anhepatic period are significantly associated with the development of PRS in LDLT.
• Meaning: These findings suggest hypothesis that in LDLT, male sex and presence of small LVEDD, large graft volume to standard liver volume ratio, and long anhepatic period are significant risk factors for PRS, and high Ca²⁺ concentration and low mean pulmonary artery pressure before reperfusion may be protective factors for PRS.

During liver transplantation (LT), postreperfusion syndrome (PRS) often occurs after portal vein reperfusion (PVR) of the liver graft. PRS was first described by Aggarwal et al¹ in 1987 as cardiovascular collapse and was defined as a >30% decrease in the mean arterial pressure below the baseline value within 5 minutes of reperfusion that persists for at least 1 minute. It has been reported that PRS was associated with numerous postoperative complications, such as graft dysfunction and renal failure, along with length of hospital stay and mortality.^2–8 The incidence of PRS in LT is 10% to 60%. Although the exact mechanism has not been elucidated, PRS can be attributed to acute acidosis, hyperkalemia, hypothermia, and abrupt release of vasodilators from the transplanted graft.^7–12

Several studies have been conducted on PRS over the years; however, controversies still exist regarding its pathophysiology, risk factors, and management.¹³ Many studies have discussed risk factors for PRS in LT, but the results are not consistent.^2–9 Additionally, most studies were performed on brain-dead donor LT but not on living donor LT (LDLT), which is frequently performed in Japan.^14–16 Furthermore, a few studies have conducted comprehensive analyses of various perioperative factors with a sufficient number of patients.

Therefore, this study aimed to identify factors associated with PRS to contribute to the anesthetic management to reduce PRS during LDLT.

METHODS

Study Population

This study is an observational and retrospective study approved by the Clinical Research Ethics Committee of Kyushu University, Fukuoka, Japan (research number: 30-385), who waived the need for individual written informed consent. The study was conducted in compliance with the principles of the Declaration of Helsinki. This manuscript adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. We included 293 patients who underwent LDLT at Kyushu University Hospital between January 2013 and September 2018. After excluding 43 patients aged <20 years, the remaining 250 patients were enrolled in the study.

Data Collection

We reviewed the medical records and anesthetic charts of 250 patients manually. Patient data included sex, age, body mass index, primary disease, general condition, comorbidities, and vital signs. Data pertaining to preoperative examination of the blood and transthoracic echocardiography, perioperative arterial blood gas (ABG) measurements, and hemodynamic parameters were collected. Data on each corresponding donor were also collected for sex, age, body mass index, blood type, graft volume (GV), and ischemia time. Almost all preoperative examinations of the blood and transthoracic echocardiography were conducted within 1 week and 1 month before surgery, respectively.

Anesthetic Management

In most cases, anesthesia was induced by intravenous administration of fentanyl, propofol, and rocuronium. All patients received an intraarterial catheter for invasive arterial pressure monitoring and blood sampling. After tracheal intubation, a pulmonary artery catheter was inserted via the internal jugular vein, and continuous cardiac output and mixed venous oxygen saturation were monitored in addition to the American Society of Anesthesiologists (ASA)-mandated monitors. Anesthesia was maintained with remifentanil combined with desflurane or isoflurane, in conjunction with a 40% to 60% oxygen-air mixture. Additional bolus infusions of fentanyl and rocuronium were provided as needed. Frequent ABG analysis guided electrolyte and acid-base management.

End Point

The primary end point was the incidence of PRS. We defined PRS as a decrease in mean arterial pressure of >30% from the baseline value, lasting at least for 1 minute within the first 5 minutes of reperfusion, regardless of the use of hypertensive drugs, as described previously.¹ Of the 250 patients, 73 (29.2%) developed PRS. Patients who developed PRS were assigned to the PRS group, and those who did not were assigned to the non-PRS group.

Measurement of Associated Factors

All data were retrospectively collected from medical records and anesthetic charts. Body mass index was calculated as weight in kilograms divided by height in meters squared. Primary diseases were categorized as viral, alcoholic, autoimmune, or other causes of hepatitis. Other causes included nonalcoholic steatohepatitis, Wilson disease, biliary atresia, drug-induced hepatitis, Caroli disease, hepatic cysts, and obstruction of the hepatic vein. Disease progression was classified as acute or chronic. The model for end-stage liver disease (MELD) score and ASA-physical status (ASA-PS) were used to evaluate the patient’s hepatic function and general condition, respectively.^17,18 ASA-PS was classified as I, II, III, or IV. A history of cardiovascular disease was defined as any episode of ischemic heart disease, valvular heart disease, cardiomyopathy, congenital heart disease, heart failure, arrhythmia, and pulmonary hypertension. The estimated glomerular filtration rate (eGFR) was calculated using the Japanese coefficient-modified chronic kidney disease epidemiology collaboration equation.¹⁹ Left ventricular wall motion on preoperative transthoracic echocardiography was evaluated to determine whether wall motion asynergy exists. ABG measurements were conducted during anesthetic induction within 1 hour before graft reperfusion and within 30 minutes after graft reperfusion in most cases, and each data point was used for analysis. Hemodynamic parameters, such as heart rate, arterial blood pressure, central venous pressure, pulmonary artery pressure (PAP), cardiac index, mixed venous oxygen saturation, and body temperature, were collected immediately before and after graft reperfusion and included in the analysis. In addition, we checked whether bolus vasopressors were administered in the minutes before or after reperfusion, and if they were used, this was deemed as bolus vasopressors being used during reperfusion.

Statistical Analysis

Patients’ baseline characteristics were compared between the PRS and non-PRS groups. Continuous variables were presented as mean values (standard deviation) or median values (interquartile range) and categorical variables as percentages. Logistic regression was used to assess the relationship between background factors and PRS and estimate odds ratios (ORs) with 95% confidence intervals for PRS. Variables that did not show a linear relationship with the logit of the outcome were transformed as restricted cubic splines. We used restricted cubic spline with 3 knots located at the 10th, 50th, and 90th percentiles of the variable distribution in the logistic regression. In the multivariable analysis, we included potential confounding factors at baseline and prereperfusion and possible factors for PRS, which indicated P <.2 in a nonadjusted model. The backward method was used to select significant variables in the multivariable analysis (P <.1 for selection criteria). Furthermore, as a sensitivity analysis, the same analysis including variables detected in the main analysis was performed, excluding those for which bolus vasopressors were used during graft reperfusion. For sample size calculation, since this was an exploratory study, we did not calculate the sample size, but we did calculate how much of a difference would be detectable based on a 29.2% incidence of PRS in a population of 250 subjects. Assuming a power of 0.8, an alpha error of 0.05, and the same binary variable for proportions, a difference of 11.8% was detectable from the current sample size. SAS software package (version 9.4; SAS Institute) was used to perform all statistical analyses. Two-sided values of P <.05 were considered statistically significant in all analyses.

RESULTS

The baseline characteristics of 250 patients included in our study are presented in Table 1. The 73 patients who developed PRS were likely to have ascites. The preoperative heart rate in the PRS group was higher than that in the non-PRS group. There were no significant differences in sex, age, liver function, general condition, comorbidities, blood pressure, or medication between the 2 groups.

Table 1. - Baseline Characteristics of Participants and OR for PRS

Variables	Non-PRS group (n = 177)	PRS group (n = 73)	OR for PRS (95% CI)	P value
Male sex, %	42.4	52.1	1.48 (0.85–2.56)	.16
Age, y	55 (10)	58 (11)	1.02 (0.996–1.05)	.11
Height, cm	161.0 (8.6)	161.0 (8.4)	0.99 (0.97–1.03)	.97
Weight, kg	62.4 (12.2)	63.2 (10.3)	1.01 (0.98–1.03)	.63
Body mass index, kg/m2	24.0 (3.7)	24.4 (3.7)	1.03 (0.96–1.11)	.42
Primary diseases, %				.06
Viral hepatitis	40.1	35.6	1.00 (reference)
Alcoholic hepatitis	13.6	17.8	1.48 (0.65–3.31)	.46
Autoimmune hepatitis	31.1	19.2	0.70 (0.33–1.44)	.03
Other causes of hepatitis	15.3	27.4	2.02 (0.97–4.22)	.04
Acute progression, %	6.8	4.1	0.59 (0.13–1.92)	.42
MELD score	14 (8)	14 (8)	1.01 (0.97–1.04)	.69
Hepatorenal syndrome, %	11.3	12.3	1.10 (0.46–2.49)	.82
Hepatopulmonary syndrome, %	4.0	4.1	1.04 (0.22–3.86)	.95
Esophageal varix, %	56.3	59.7	1.15 (0.66–2.02)	.62
Ascites, %	37.3	53.4	1.93 (1.11–3.36)	.02
History of abdominal surgery, %	31.1	21.9	0.62 (0.32–1.16)	.15
ASA-PS, %				.21
I or II	9.6	13.7	1.00 (reference)
III	80.2	69.9	0.61 (0.27–1.46)	.08
IV	10.2	16.44	1.13 (0.39–3.35)	.39
Hypertension, %	22.6	21.9	0.96 (0.49–1.83)	.91
SBP, mm Hg	115 (17)	116 (16)	1.00 (0.99–1.02)	.66
DBP, mm Hg	66 (12)	65 (12)	1.00 (0.97–1.02)	.76
HR, bpm	79 (14)	83 (15)	1.02 (1.00–1.04)	.04
Diabetes, %	20.9	24.7	1.24 (0.64–2.34)	.52
Dyslipidemia, %	2.8	4.1	1.47 (0.30–6.17)	.60
History of cardiovascular diseases, %	6.8	5.5	0.80 (0.22–2.38)	.70
Dialysis, %	4.5	6.9	1.55 (0.46–4.83)	.45
Medication, %
CCB	7.3	4.1	0.54 (0.12–1.74)	.35
ARB	6.2	8.2	1.35 (0.45–3.70)	.57
ACE-I	0.6	1.4	2.44 (0.10–62.32)	.53
β-blocker	5.1	8.2	1.67 (0.54–4.82)	.35
Diuretic	44.6	41.1	0.87 (0.50–1.50)	.61
Anticoagulant or antiplatelet drug	4.0	5.5	1.41 (0.36–4.82)	.59

Data are presented as mean values (standard deviation) or percentages. Logistic regression was used to estimate ORs with 95% CIs for PRS.

Abbreviations: ACE-I, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ASA-PS, American Society of Anesthesiologists Physical Status; CCB, calcium channel blocker; CI, confidence interval; DBP, diastolic blood pressure; HR, heart rate; MELD, model for end-stage liver disease; OR, odds ratio; PRS, postreperfusion syndrome; SBP, systolic blood pressure.

The preoperative findings of blood examination and transthoracic echocardiography are presented in Table 2 and Supplemental Digital Content, Table 1, https://links.lww.com/AA/D909. Left ventricular end-diastolic diameter (LVEDD) and left atrial diameter in the PRS group were smaller than those in the non-PRS group, whereas other factors were not statistically different between the 2 groups.

Table 2. - Preoperative Examination of the Arterial Blood Gas and Transthoracic Echocardiography and OR for PRS

Variables	Non-PRS group (n = 177)	PRS group (n = 73)	OR for PRS (95% CI)	P value
Arterial blood gas
pH	7.44 (0.06)	7.43 (0.06)	0.15 (0.002–11.76)	.39
HCO3 –, mmol/L	25.0 (3.3)	25.1 (3.0)	1.01 (0.93–1.10)	.80
BE, mmol/L	0.5 (3.7)	0.6 (3.4)	1.00 (0.93–1.08)	.96
Lactic acid, mg/dL	11 (9–14)	12 (10–15)	1.01 (0.98–1.04)	.61
Transthoracic echocardiography
LVEDD, mm	48.8 (5.6)	46.5 (6.0)	0.93 (0.89–0.98)	.006
LVESD, mm	28.4 (4.6)	27.8 (5.0)	0.97 (0.91–1.03)	.36
IVST, mm	8.8 (2.0)	8.5 (1.2)	0.90 (0.73–1.07)	.28
LVPWT, mm	8.9 (1.3)	8.7 (1.3)	0.92 (0.74–1.13)	.43
EF, %	72.3 (6.3)	71.5 (5.5)	0.98 (0.93–1.03)	.36
LAD, mm	40.6 (7.7)	37.8 (6.4)	0.94 (0.90–0.98)	.007
Mitral E/E’	11.1 (3.5)	10.9 (3.9)	0.98 (0.90–1.07)	.70
LVWM asynergy, %	4.62	4.11	0.88 (0.19–3.16)	.86

Data are presented as mean values (standard deviation), median values (interquartile range), or percentages. Logistic regression was used to estimate ORs with 95% CIs for PRS.

Abbreviations: BE, base excess; CI, confidence interval; EF, ejection fraction; HCO₃⁻, bicarbonate ion; IVST, interventricular septum thickness; LAD, left atrial diameter; LVEDD, left ventricular end-diastolic diameter; LVESD, left ventricular end-diastolic diameter; LVPWT, left ventricular posterior wall thickness; LVWM, left ventricular wall motion; OR, odds ratio; pH, power of hydrogen; PRS, postreperfusion syndrome.

ABG measurements and hemodynamic parameters before graft reperfusion are presented in Table 3 and Supplemental Digital Content, Table 2, https://links.lww.com/AA/D909. The findings of ABG, hemocytes, and coagulation function before graft reperfusion were not significantly different between the 2 groups. Systolic, diastolic, and mean arterial pressures; systolic and mean PAPs; cardiac index; and mixed venous blood oxygen saturation before graft reperfusion were higher in the PRS group than in the non-PRS group. Bolus vasopressors were more commonly used in the PRS group during graft reperfusion. Core temperature before reperfusion was significantly higher in the non-PRS group than in the PRS group.

Table 3. - Arterial Blood Gas Measurement and Examination of the Blood Before Reperfusion and OR for PRS

Variables	Non-PRS group (n = 177)	PRS group (n = 73)	OR for PRS (95% CI)	P value
Arterial blood gas before reperfusion
pH	7.37 (0.06)	7.36 (0.06)	0.02 (<0.001–1.89)	.09
Paco 2, mm Hg	36.8 (4.0)	37.1 (2.9)	1.02 (0.95–1.10)	.56
Pao 2, mm Hg	205.3 (50.7)	214.3 (48.8)	1.00 (0.998–1.01)	.21
Hco 3 –, mmol/L	21.7 (2.6)	21.3 (2.4)	0.93 (0.83–1.04)	.19
BE, mmol/L	−3.4 (3.0)	−3.9 (2.9)	0.94 (0.86–1.03)	.21
Na+, mmol/L	140.1 (4.2)	141.0 (4.6)	1.05 (0.98–1.12)	.15
K+, mmol/L	3.8 (0.7)	3.8 (0.5)	0.91 (0.59–1.36)	.65
Cl–, mmol/L	107.3 (4.7)	107.4 (4.5)	1.00 (0.95–1.07)	.89
Ca2+, mmol/L	1.09 (0.18)	1.05 (0.15)	0.21 (0.04–1.09)	.07
Mg2+, mmol/L	0.47 (0.10)	0.48 (0.09)	2.04 (0.11–33.60)	.62
Lactic acid, mg/dL	30 (24–40)	33 (25–43)	1.01 (0.995–1.03)	.16
Glucose, mg/dL	158 (34)	154 (36)	1.00 (0.99–1.004)	.39
Hemocytes before reperfusion
Hb, g/dL	9.0 (1.6)	8.9 (1.5)	0.96 (0.81–1.15)	.66
Hct, %	27.2 (4.7)	26.8 (4.3)	0.98 (0.92–1.04)	.52
WBC, ×103/µL	3.10 (2.20–4.40)	3.60 (2.20–4.20)	1.02 (0.89–1.15)	.80
Plt, ×104/µL	5.1 (3.6–7.0)	5.4 (3.6–6.8)	0.97 (0.61–1.55)	.90
Coagulation function before reperfusion
PT-INR	1.54 (0.34)	1.50 (0.27)	0.64 (0.21–1.59)	.38

Data are presented as mean values (standard deviation) or median values (interquartile range). Logistic regression was used to estimate ORs with 95% CIs for PRS.

Abbreviations: BE, base excess; Ca²⁺, calcium ion; CI, confidence interval; Cl⁻, chloride ion; Hb, hemoglobin; HCO₃⁻, bicarbonate ion; Hct, hematocrit; K⁺, potassium ion; Mg²⁺, magnesium ion; Na⁺, sodium ion; OR, odds ratio; Paco₂, partial pressure of carbon dioxide; Pao₂, partial pressure of oxygen; pH, power of hydrogen; Plt, platelet; PRS, postreperfusion syndrome; PT-INR, prothrombin time-international normalized ratio; WBC, white blood cell.

The intraoperative and donor data are presented in Table 4. The anhepatic period was significantly longer in the PRS group than in the non-PRS group. GV and the ratio of GV to the standard liver volume (GV/SLV) were larger in the PRS group than in the non-PRS group. The compatibility of the recipient and donor blood types and graft ischemic time were not significantly different between the 2 groups.

Table 4. - Intraoperative and Donor Information and OR for PRS

Variables	Non-PRS group (n = 177)	PRS group (n = 73)	OR for PRS (95% CI)	P value
Intraoperative information
Operation time, min	717 (130)	706 (136)	1.00 (0.997–1.001)	.56
Anesthesia time, min	848 (135)	836 (136)	1.00 (0.997–1.001)	.51
Venous bypass, %	1.69	2.74	1.63 (0.21–10.06)	.60
Anhepatic period, min	113 (63–185)	166 (102–216)	1.005 (1.001–1.008)	.006
Donor information
Male sex, %	65	56.2	0.69 (0.40–1.21)	.19
Age, y	39 (11)	38 (10)	0.99 (0.97–1.02)	.47
Body mass index, kg/m2	22.2 (2.5)	21.9 (2.6)	0.96 (0.86–1.07)	.46
Blood type, %				.98
Matched	53.7	54.8	1.00 (reference)
Unmatched	22.0	21.9	0.97 (0.48–1.92)	.99
Incompatible	24.3	23.3	0.94 (0.47–1.82)	.88
GV, g	484 (116)	535 (123)	1.004 (1.001–1.01)	.003
GV/SLV, %	41.2 (8.7)	46.2 (10.2)	1.06 (1.03–1.09)	<.001
Graft ischemic time, min	152 (118–205)	164 (123–217)	1.00 (0.998–1.01)	.35

Data are presented as mean values (standard deviation), median values (interquartile range), or percentages. Logistic regression was used to estimate odds ratios with 95% confidence intervals ORs with 95% CIs for PRS. 
Abbreviations: CI, confidence interval; GV, graft volume; GV/SLV, graft volume to standard liver volume ratio; OR, odds ratio; PRS, postreperfusion syndrome.

Multivariable-adjusted logistic regression with the backward method was performed for all variables with P <.2 in a nonadjusted model, and significant factors associated with PRS were detected (Table 5). The multivariable-adjusted OR for PRS was significantly higher in men than in women. Small LVEDD and low calcium ion (Ca²⁺) concentration before reperfusion were significantly associated with developing PRS. Mean PAP before reperfusion, anhepatic period, and GV/SLV, included as restricted cubic spline, were also significantly associated with PRS. The coefficient of determination in the final model was R² = 0.256, and approximately 25.6% of the variation in the incidence of PRS could be explained by these factors.

Table 5. - Multivariable-Adjusted ORs for PRS After Variable Selection

Variables	OR for PRS (95% CI)	P value
Men (versus women)	2.45 (1.26–4.75)	.008
LVEDD (per 1-mm increase)	0.90 (0.85–0.95)	<.001
Ca2+ before reperfusion (per 0.1-mmol/L increase)	0.74 (0.60–0.91)	<.001
Mean PAP before reperfusion	Not available	.003
Anhepatic period	Not available	.02
GV/SLV	Not available	.03

Multivariable adjustment was made for covariates that showed P <.2 in a nonadjusted model, and variable selection was made by the backward method (P <.1 for selection criteria). Logistic regression was used to estimate associations with PRS, and mean PAP, anhepatic period, and GV/SLV, which did not show a linear relationship with the logit of the outcome, were transformed as restricted cubic splines.

Abbreviations: Ca²⁺, calcium ion; CI, confidence interval; GV/SLV, graft volume to standard liver volume ratio; LVEDD, left ventricular end-diastolic diameter; OR, odds ratio; PAP, pulmonary artery pressure; PRS, postreperfusion syndrome.

The relationship between each associated factor and the incidence of PRS is presented in the Figure. LVEDD and Ca²⁺ concentration showed a negative linear relationship with the incidence of PRS, and the anhepatic period and GV/SLV had a positive almost linear relationship with the incidence of PRS. Contrastingly, the incidence of PRS decreased as mean PAP fell below approximately 20 mm Hg.

Figure.: Relationship between each risk factor and incidence of PRS. The relationships between LVEDD (A), Ca²⁺ level (B), mean PAP (C), anhepatic period (D), and GV/SLV (E), with predicted values of PRS incidence. Logistic regression was used to calculate the predicted PRS incidence, and mean PAP, anhepatic period, and GV/SLV, which did not show a linear relationship with the logit of the outcome, were transformed as restricted cubic spline and draw spline curves for each variable. The blue ranges indicate 95% confidence intervals. Ca²⁺ indicates calcium ion; CI, confidence interval; GV/SLV, graft volume ratio to standard liver volume; LVEDD, left ventricular end-diastolic diameter; PAP, pulmonary artery pressure; PRS, postreperfusion syndrome.

The results were similar in a sensitivity analysis that excluded studies in which bolus vasopressors were used during graft reperfusion (Supplemental Digital Content, Table 3, https://links.lww.com/AA/D909).

DISCUSSION

We found that male sex and presence of small LVEDD, low Ca²⁺ concentration and high mean PAP before reperfusion, long anhepatic period, and high GV/SLV were significantly associated with the development of PRS. Furthermore, although our data do not support causal inferences, the data suggest that high Ca²⁺ concentration and low mean PAP before reperfusion might be protective factors for PRS. Since Ca²⁺ concentration and mean PAP are adjustable factors in anesthetic management, they may play an important role in avoiding the adverse effects of PRS.

Previous retrospective studies on brain-dead donor LT have reported factors associated with PRS, namely, donor age, Child-Pugh score, MELD score, and cold ischemia time, albeit inconsistently.^{3,6–9,12,20–22} Intriguingly, these factors in LDLT in our study are completely different from those reported in studies with brain-dead donor LT; however, the reason for this discrepancy cannot be clarified and may be owing to the difference in patients. In contrast, among the 3 previous retrospective studies examining PRS association in LDLT,^14–16 1 study reported that intraoperative calcium supplementation reduced PRS incidence,¹⁴ supporting the results of our study. However, sex^14–16 and anhepatic period^15,16 were not associated with PRS in those previous studies, which might be owing to racial differences and the small sample size of the previous studies. Other factors, including preoperative left ventricular diameter, PAP before reperfusion, and GV, have not been examined in previous studies, which makes our study novel. However, as this is a retrospective observational study, we could not address the causal relationship between the results obtained. Nevertheless, our study may serve as a catalyst for clinical strategies in LDLT to reduce PRS.

In our study, men undergoing LDLT were associated with significantly increased risk of PRS, possibly owing to the difference in hormones^23,24; however, the exact mechanism is unknown. A small LVEDD was associated with developing PRS. It could be hypothesized that a smaller left ventricle tends to be more easily affected by an abrupt increase in preload during reperfusion, which causes a temporary state of left ventricular failure and subsequent hypotension.^25,26 However, this is an inference of the association between LVEDD and PRS from a multivariable analysis that excluded various confounding factors. Considering that the actual difference between the 2 groups was about 2 mm, which is clinically negligible, the association between LVEDD and PRS should be considered not only in terms of LVEDD alone but also in terms of other factors and should be interpreted with caution.

Additionally, a high mean PAP before reperfusion was associated with developing PRS. This could be partly explained by the same mechanism mentioned above; when the mean PAP is high, a large volume can already exist in the left ventricle, and during reperfusion, excess preload could induce left ventricular failure. However, caution should be exercised when interpreting the association between PAP and PRS because PAP is influenced not only by left ventricular preload but also by pulmonary vascular resistance. Even when the mean PAP is high, PRS might be less likely to occur because it is harder to pump blood from the right ventricle to the left. Therefore, a high mean PAP may not have been associated with developing PRS to the extent to which it did in our study. However, the relationship between mean PAP and PRS may also be related to right heart failure. A high mean PAP indicates a preliminary stage of right heart failure, and graft reperfusion may have caused an acute right heart failure and acute decrease in arterial pressure.²⁷

Low Ca²⁺ levels were associated with developing PRS in our study. This is not surprising because Ca²⁺ is indispensable for both systolic²⁸ and diastolic²⁹ functions of the heart. With sufficient Ca²⁺ level, blood pressure is unlikely to fall because the blood can be pumped by sufficient cardiac contraction and expansion to counteract the excessive increase in preload caused by reperfusion.

A large GV and long anhepatic period were also associated with developing PRS. The reason for this could be explained by the release of large amounts of vasodilators during reperfusion,^30–32 although the pathophysiology of PRS has not been fully elucidated.

The strength of our study is that we were able to examine the association between various perioperative factors and PRS in a sufficient number of patients. However, our study has several limitations. First, as this is a retrospective study, anesthetic management, including the use of vasopressors, was not standardized. Therefore, misclassification of PRS may have occurred because, in some cases, attending anesthesiologists have used vasopressors to prevent PRS during reperfusion. Although the results were similar even when we excluded studies that used bolus vasopressors during reperfusion, we could not include this factor in the multivariable analysis because it was not possible to distinguish whether bolus vasopressors were administered before or after reperfusion. Second, as this was an observational study, we could not address causation. Further studies, such as randomized controlled trials, are required to prove a causal relationship. Third, it is difficult to generalize the results to other racial groups because this is a single-center study conducted in Japan. Fourth, although we included various factors in analysis, we were unable to adequately control for confounders in the final model because the variable selection analysis may have excluded relevant confounders of PRS. This analysis is exploratory in nature and generated hypotheses about relationships but did not provide conclusive evidence about those associations. Fifth, because we considered many variables, type I error may have occurred.

In conclusion, male sex and presence of small LVEDD, large GV/SLV, low Ca²⁺ concentration and high PAP before reperfusion, and long anhepatic period were significantly associated with PRS in LDLT. These findings may be useful for predicting the probability of PRS in patients undergoing LDLT and proposing an effective intervention in anesthetic management. Further investigations, such as randomized controlled trials, are needed to establish more detailed strategies for preventing PRS in LDLT.

ACKNOWLEDGMENTS

We thank Editage (www.editage.com) for English language editing.

DISCLOSURES

Name: Kaoru Umehara, MD, PhD.

Contribution: This author helped conceive, design, acquire, analyze, and interpret the study data, draft the manuscript, provide critical revisions, and approve the final manuscript.

Name: Yuji Karashima, MD, PhD.

Contribution: This author helped interpret the study data, provide critical revisions, and approve the final manuscript.

Name: Tomoharu Yoshizumi, MD, PhD.

Contribution: This author helped interpret the study data, provide critical revisions, and approve the final manuscript.

Name: Ken Yamaura, MD, PhD.

Contribution: This author helped interpret the study data, provide critical revisions, and approve the final manuscript.

This manuscript was handled by: Tong J. Gan, MD.