Mannitol for Prevention of Acute Kidney Injury After Liver Transplantation

A Randomized Controlled Trial

Moataz Maher Emara; Doaa Galal Diab; Amr Mohamed Yassen; Maha A. Abo-Zeid


BMC Anesthesiol. 2022;22(393) 

In This Article


This is a triple-blinded randomized controlled study of the role of mannitol in the prevention of AKI in LDLT. The AKI incidence was almost similar between the groups. Most intraoperative hemodynamic parameters did not show statistical differences between the groups. Both Na+ and Cl- were higher in the S group at 5 min before and after portal vein declamping (graft reperfusion). The intraoperative UO was higher in the M group.

Patient characteristics and AKI risk factors including operative time, cold ischemia, warm ischemia, anhepatic time, intraoperative blood transfusion, and the incidence of PRS were comparable in both groups.

Mannitol could not prevent early AKI following LDLT with a tendency toward a higher AKI incidence in the M group—11/39 (28.2%) versus 9/41 (22%) in the S group, respectively (P-value = 0.518).

Mannitol did not prevent acute renal failure during LT which is in agreement with a study by Whitta et al. Nonetheless, they only studied 25 patients with 12 cases in the M group.[9] However, they started mannitol infusion after induction of anesthesia, which is a long time before reperfusion while the plasma elimination half-life of mannitol was 2.44 h and the duration of action may extend to 8 h after the end of infusion.[14] Mannitol could not prevent the effects of hepatic ischemic-reperfusion injury on renal function unlike in the biochemical and animal studies.[5,15,16] Mannitol could not downstage the AKI even when the AKI occurred, as shown in the existing study. There were no statistical or clinical differences between the groups regarding the AKI stages.

Mannitol did not reduce the incidence of PRS in the current study. The PRS incidence is 29/39 (74.4%) and 31/41 (75.6%) in M and S groups, respectively, P-value = 0.897. Mannitol infusion during the anhepatic phase improved the PRS during LDLT unlike Shameddini et al. In contrast, they did not calculate the incidence of PRS, but they calculated the differences in MAP and CO before and after portal vein declamping, which is statistically questionable.[8]

Total intraoperative UO and the anhepatic UO were increased in the M group. This is easily explained by mannitol's diuretic effect. The increased UO in the M group did not imply preserved renal function post-LDLT.[14]

Most intraoperative hemodynamic changes were comparable between the groups. The SVR was higher in the M group at 5 min after hepatic declamping and at closure time. The explanation is not clear, but this resulted in a higher CI in the S group at the closure time.

This is contradictory to the study of Chatterjee et al.[17] which showed a significant decrease in SVR, at 5 min and 15 min after infusion of the same mannitol dose. This difference may be explained by the specific criteria of LDLT as there was widespread use of norepinephrine, epinephrine, and furosemide which increases SVR.

The K+ and ionized Ca+2 did not show differences between both groups regarding the intraoperative electrolyte changes. However, both [Na+] and [Cl-] were significantly lower in the M group 5 min before and 5 min after portal reperfusion. Those changes were transient. This could be interpreted by mannitol's volume increase effects which resulted in the dilution of both Na+ and Cl-.[14] Osmolality changes due to mannitol may also result in Na+ and Cl- loss to compensate for the increased osmolality.[18]

The relative constant ionized [Ca+2] and [K+] in our study may be due to the close monitoring, continuous infusion, and correction of calcium chloride and potassium throughout the operation, especially around the reperfusion phase.

The pH, serum lactate, and LDH in the first two ICU days were similar between both groups regarding the liver graft function. Postoperative INR and albumin did not show differences except for albumin at 3 months post-LT, where albumin was statistically higher in the mannitol group. However, this difference is not clinically different as the average albumin level was in the normal range in both groups, 4.44 g/dl in the M group vs. 4.21 g/dl in the S group.

Liver enzymes (AST and ALT), total bilirubin, and GGT did not show differences between both groups except serum bilirubin on the 7th postoperative day. The 7th-day total serum bilirubin was higher in the M group with a median of 4.7 mg/dl versus 2 mg/dl in the S group. Nevertheless, bilirubin normalized on the 28th day and 3rd-month post-LT.

Likewise, the duration of ICU stay, the incidence of postoperative surgical complications, and the 3-month survival period were all similar between the groups.

We used serum creatinine for the diagnosis and grading of AKI despite its limitations in renal function evaluation. Muscle wasting and ascites besides hyperbilirubinemia overestimate the S.Cr measurement.[19,20]

Some authors suggested a corrected S.Cr formula to compensate for the acute fluid change overload in the intraoperative and intensive care period.[21] However, this formula was not validated, especially in liver transplant or cirrhotic patients. Furthermore, guidelines still consider the S.Cr as the most valid marker for AKI.[22–24]

The definition of AKI in our study follows the International Club of Ascites revised definition of KDIGO guidelines, defined as a 0.3 mg/dl increase in the serum creatinine (S.Cr) in the early 48 postoperative hours.[11,12] They removed the UO criteria from the definition as cirrhotic patients may have oliguria, but still have normal renal function. The International Club of Ascites can define both early and late AKI. Early refers to the AKI within 48 h and late refers to the AKI within 7 days.[11,12]

We did not use new markers detecting AKI as all guidelines still use S.Cr as the standard marker. No specific marker has been well-validated in LT. Our study has limitations: the study was not pre-registered, but we strictly followed the pre-designed protocol, approved by the IRB and is uploaded with the manuscript; the incidence of early AKI was lower than in the pilot study was higher than in what found in the study which makes the sample size questionable. We think that the enhanced fluid management during the study period reduced the incidence of AKI in this cohort. The study sample size (80 cases) is large when compared to similar studies in LDLT; however, it was insufficient to detect mannitol effects on the renal and liver graft functions.

In conclusion, the current LDLT recipient sample was insufficient to demonstrate that pre-reperfusion 1 g/kg mannitol infusion would reduce the risk of early AKI, PRS, early postoperative graft function, or the 3-month survival period.