A randomized clinical trial testing the anti-inflammatory effects of preemptive inhaled nitric oxide in human liver transplantation

Abstract

Decreases in endothelial nitric oxide synthase derived nitric oxide (NO) production during liver transplantation promotes injury. We hypothesized that preemptive inhaled NO (iNO) would improve allograft function (primary) and reduce complications post-transplantation (secondary). Patients at two university centers (Center A and B) were randomized to receive placebo (n = 20/center) or iNO (80 ppm, n = 20/center) during the operative phase of liver transplantation. Data were analyzed at set intervals for up to 9-months post-transplantation and compared between groups. Patient characteristics and outcomes were examined with the Mann-Whitney U test, Student t-test, logistic regression, repeated measures ANOVA, and Cox proportional hazards models. Combined and site stratified analyses were performed. MELD scores were significantly higher at Center B (22.5 vs. 19.5, p<0.0001), surgical times were greater at Center B (7.7 vs. 4.5 hrs, p<0.001) and warm ischemia times were greater at Center B (95.4 vs. 69.7 min, p<0.0001). No adverse metabolic or hematologic effects from iNO occurred. iNO enhanced allograft function indexed by liver function tests (Center B, p<0.05; and p<0.03 for ALT with center data combined) and reduced complications at 9-months (Center A and B, p = 0.0062, OR = 0.15, 95% CI (0.04, 0.59)). ICU (p = 0.47) and hospital length of stay (p = 0.49) were not decreased. iNO increased concentrations of nitrate (p<0.001), nitrite (p<0.001) and nitrosylhemoglobin (p<0.001), with nitrite being postulated as a protective mechanism. Mean costs of iNO were $1,020 per transplant. iNO was safe and improved allograft function at one center and trended toward improving allograft function at the other. ClinicalTrials.gov with registry number 00582010 and the following URL:http://clinicaltrials.gov/show/NCT00582010.

Figure 1. CONSORT diagram for center A.

Assessment of Eligibility and Inclusion in the Inhaled Nitric Oxide for Attenuating Liver Allograft Injury Study at Center A.

Figure 2. CONSORT diagram for center B.

Assessment of Eligibility and Inclusion in the Inhaled Nitric Oxide for Attenuating Liver Allograft Injury Study at Center B.

Figure 3. Patient demographic differences between study sites.

Panels A–E show respectively patient weight, MELD scores, surgery, warm ischemic times and donor risk index (DRI). Each data point represents individual patient (n = 37–40 for Center A, and n = 40 for Center B). Indicated P-values are from unpaired t-test (panel A and E) or Mann-Whitney U test (panels B–D).

Figure 4. Changes in metHb and nitrogen dioxide (NO₂) as a function of blood draw (BD) in placebo (○) or iNO (▪) groups.

Data are mean ± SEM (n = 20 for each group, except UAB placebo where n = 19 and UW iNO where n = 17–20. *P<0.0001 by 2-way ANOVA with Bonferroni multiple comparisons post-test.

Figure 5. iNO therapy did not affect volume of packed RBC (pRBC) or platelets transfused during liver transplant surgery at either study site.

Data are mean ± SEM (n = 20 for each group, except UAB placebo where n = 19). Indicated p-values calculated by unpaired t-test.

Figure 6. Center dependent effects of iNO therapy on post-transplant liver function.

Liver function post-transplantation was assessed by following post-surgery time-dependent changes in serum AST, serum ALT, PT time, alkaline phosphatase and bilirubin as indicated. Panels A,B, C shows respectively data for Center A, Center B and the two centers combined. Data represented by symbols (□ = placebo, • = iNO) are presented as percent change relative to first measurement post-surgery and are mean ± SEM (n = 20, except Center A placebo where n = 19; n = 39–40 for combined). Fitted lines (solid line = placebo, dotted line = iNO) show means after adjustment for MELD, cold ischemic time, weight, BMI (>25), warm ischemia time (WIT) and donor age. P-values indicate significance of iNO vs. placebo for adjusted data. Absolute values for LFT parameters measured immediately (<1 h) post-surgery for were: AST 722.8±114 U/L (placebo) and 636.1±136.9 U/L (iNO) at Center A; 1432.1±222.5 U/L (placebo) and 1429.1±134.3 U/L (iNO) at Center B; ALT 479.5±67.1 U/L (placebo) and 423.7±77.8 U/L (iNO) at Center A; 904.3±155.4 U/L (placebo) and 987.7±109.2 U/L (iNO) at Center B; PT time 31.9±2.4 s (placebo) and 31.7±2 s (iNO) at Center A; 22.6±0.8 s (placebo) and 24.6±0.8 s (iNO) at Center B; Alk Phos 77.7±14.3 U/L (placebo) and 70.1±8.9 U/L (iNO) at Center A; 57.2±4.4 U/L (placebo) and 68.4±13 U/L (iNO) at Center B; Bili-total 3.6±0.4 (placebo) and 3.66±0.33 (iNO) at Center A; 3.2±0.5 (placebo) and 4.2±0.5 U/L (iNO) at Center B. Data are normalized to ALT and AST levels measured immediately (<1 h) post-surgery and were 601.8±145.4 U/L (placebo) and 689.3±149.5 U/L (iNO) for ALT and 922.1±228.7 U/L (placebo) and 940.9±211.3 U/L (iNO) for AST.

Figure 7. iNO therapy did not affect hospital or ICU length of stay.

Kaplan-Meier plots shown and with P-values from Cox proportional hazards models adjusting for MELD, cold ischemic time, weight, BMI, warm ischemic time or donor age within each center. Placebo = ○, iNO = ▪.

Figure 8. Rates of post-operative hepatobiliary complications.

A. Data are presented as ‘no’ or ‘yes’ indicating the absence or presence of hepatobiliary complications respectively within 9 months of surgery.

Figure 9. Effects of iNO on reperfusion induced injury and PMN accumulation.

Liver biopsy samples were collected pre- (LB1, □) and 1 hr post dual reperfusion (LB2, ▪) and assessed for injury by histopathologic evaluation (Panel A, B) and infiltration of PMN (panel E,F). The increase (LB2 – LB1) in pathology score and PMN infiltration are shown in panels C–D (center A) and G–H (Center B), respectively. P-values indicated on graph are by paired t-test for panels A–B or by * unpaired t-test or ^#Mann-Whitney U (panels C–D). Panels A, C, E, G from Center A cohort. Panels B ,D, F, H are from the Center B cohort.

Figure 10. Effects of iNO on reperfusion induced injury and PMN accumulation.

Liver biopsy samples were collected pre- (LB1, □) and 1 hr post dual reperfusion (LB2, ▪) and assessed for injury by histopathologic evaluation (Panel A, B) and infiltration of PMN (panel C, D). Data from both UAB and UW cohorts are combined (n = 38). P-values indicated on graph are by paired t-test for panels A and C or by * unpaired t-test (panels B and D).

Figure 11. Effects of iNO on TUNEL staining pre- and post-reperfusion.

Hepatic cell death was assessed by TUNEL staining in central vein and triad regions pre- (LB1) and 1 hr post-dual reperfusion (LB2) in placebo and iNO treated patients. Panel A and B show representative immunofluorescence images for TUNEL staining (green = TUNEL positive cell, blue is Hoechst 33342 staining of nuclei, magnification 40×). Panel C and D compare the magnitude of increase in TUNEL staining between placebo and iNO groups for Center A and Center B cohorts respectively. Indicated P-values calculated by Mann-Whitney U-test.

Figure 12. Individual responses to iNO on TUNEL staining pre- and post-reperfusion.

Panel A and B show changes in TUNEL positive nuclei in hepatic central vein area and panels C and D in the hepatic triad area for Center A (panel A and B) and Center B (panel C and D) cohorts. Indicated P-values determined by Wilcoxon rank-sum (n = 18–20).

Figure 13. Effects of iNO on NO-derived metabolites.

Plasma nitrate, nitrite and RBC FeNO (nitrosylhemoglobin) was measured in paired arterial and venous samples collected at different times during the intraoperative period (□ = placebo; • = iNO). The first time point denotes pre- initiation of placebo or iNO. The third and fourth data points represent pre- and 1 hr post-reperfusion samplings. Data show mean ± SEM (n = 19–20). P<0.001 by 2-way ANOVA for all placebo vs. iNO comparisons (except FeNO in arterial group for Center B).

Figure 14. Effects of iNO on plasma and RBC nitroso species.

Plasma nitroso (Panel A) or RBC nitroso (Panel B) were measured and normalized to protein and heme, respectively. Data show measurements from paired arterial and venous samples collected at different times during the intraoperative period (□ = placebo; • = iNO). The first time point denotes pre-initiation of placebo or iNO. The third and fourth data points represent pre- and 1 h post-reperfusion samplings. Data show mean ± SEM (n = 19–20).

Figure 15. Effects of iNO on liver nitrite levels pre- (LB1) and post-reperfusion (LB2).

Nitrite was measured in paired liver biopsies and data normalized to protein. No significant differences in nitrite levels were observed pre- vs. post-reperfusion or between placebo and iNO treatments.

Figure 16. iNO effects on plasma ceruloplasmin activity.

Plasma ceruloplasmin activity was measured in venous samples collected during surgery from patients administered placebo (□) or iNO (•). Data show mean ± SEM (n = 19–20).