Objectives: Postoperative hepatic artery resistive index is strongly influenced by perioperative hemodynamics and graft-related factors. However, evidence linking prognostic variables to early postoperative hepatic artery resistive index remains limited and hete-rogeneous. This study explored changes in hepatic artery resistive index and whether changes were associated with patient-level exposures and prog-nostic variables.
Materials and Methods: We analyzed retrospective data from 323 patients who underwent deceased donor liver transplant at the Montaserieh Transplant Center Registry of Mashhad University of Medical Sciences between 2014 and 2021. Patient demographic charac-teristics and clinical assessments were analyzed. Doppler sonographies performed before and during the first 3 days posttransplant were used to measure hepatic artery resistive index. We used SPSS version 27 and R-Studio version 4.5.1 for analyses.
Results: Mean age of patients was 48.71 ± 1.022 years. Mean hepatic artery resistive index increased among patients aged ≤50 years from day 0 (0.762 ± 0.149) to day 1 (0.789 ± 0.141), showed continuous upward trend in patients aged >50 years, increasing from day 0 (0.778 ± 0.136) to day 3 (0.835 ± 0.105), and was significantly higher in patients aged >50 years on day 2 (0.031; P = .012) and day 3 (0.044; P < .001) compared with patients aged ≤50 years. Older deceased donor age was associated with increased mean hepatic artery resistive index over time (0.083, P = .029). Although baseline resistive index and liver injury markers were significantly associated with subsequent changes in hepatic artery resistive index, no significant association was observed for Model for End-Stage Liver Disease score.
Conclusions: With an association shown between age and higher hepatic artery resistive index, trend analysis may help identify high-risk groups for arterial thrombosis. Further research is needed to better clarify the relationships between Model for End-Stage Liver Disease scores and clinical outcomes.

Key words : Generalized estimation equation, Hepatic artery thrombosis, Liver transplant, Longitudinal analysis

Introduction

Liver transplantation (LT) is a remarkably effective and distinctive treatment for end-stage liver disease that substantially improves patient survival. Nevertheless, as a major and invasive surgical procedure, LT inherently poses risks of potentially life-threatening complications.¹
Hepatic artery thrombosis is a serious comp-lication that jeopardizes graft survival and often necessitates retransplant in nearly half of affected patients. In adults, this condition is associated with a high mortality rate of 35%.^2-4 When thrombosis is detected early after LT, hepatic artery revascula-rization may be attempted through surgical throm-bectomy and intra-arterial thrombolysis.³
Despite advances in surgical methods, immuno-logy, and medical science aimed at reducing life-threatening complications, hepatic artery thrombosis remains a serious challenge and greatly increases the risk of death after surgery. If revascularization is not performed quickly, risk of mortality exceeds 50%.⁵ In the absence of active surgical tactics, graft loss is observed in more than 80% of cases of hepatic artery thrombosis.⁶ Hepatic artery thrombosis is charac-terized by the absence of color and spectral Doppler signals along the course of the hepatic artery.⁷
Duplex Doppler sonography is a noninvasive technique that has been widely used in clinical settings over the past decades to assess hepatic blood flow. Ultrasonography provides several advantages, inclu-ding ease of use, safety, accuracy, and noninvasi-veness. The ability of ultrasonography to deliver real-time, dynamic, and accurate assessments makes this a diagnostic method of choice for evaluation of patients after LT.⁸
Hepatic artery resistive index (HARI) is the most useful and practical Doppler parameter for evaluation of hepatic artery blood flow over time. As a relative measure, HARI reflects both systolic and diastolic flow, making it ideal for dynamic monitoring.⁶ Measure-ments of HARI obtained at duplex Doppler sonog-raphy provides a semiquantitative estimate of the resistance to arterial blood flow within the liver.⁹ Hepatic artery resistive index values are considered normal when they are between 0.55 and 0.79 and high when they are between 0.80 and 0.89 and exceeding 0.90.6 Instantaneous increases in the resistance index (HARI >0.8) often occur within 2 days after transplant, followed by a subsequent return to normal hepatic arterial parameters.¹⁰
In most cases of increased HARI, HARI recovers to the normal range without graft complications. In general, the increase of HARI has been associated with various factors, such as hepatic artery spasm,¹¹ graft edema, or prolonged duration of cold ischemia. Moreover, increased HARI is more frequently obser-ved in LT with older recipients.¹²
In a study from Shabunin and colleagues that evaluated the negative dynamics of HARI as a predictor of early thrombosis in LT recipients, patients with normal or elevated HARI on the first day posttransplant, followed by an increase during observation, had an 18.1% incidence of arterial thrombosis.⁶ In contrast, patients with an extremely high HARI on the first day posttransplant had an 11.1% incidence. Additional risk factors for throm-bosis included recipient and donor age of >50 years.⁶ However, most studies examining HARI after LT have neither clearly defined trends in HARI nor explored the potential influence of Model for End-Stage Liver Disease (MELD) scores on these trends.
The MELD score has been shown to be an accurate predictor of survival in patients with liver disease who have not undergone transplant. Recent US investigations have shown that MELD score obtained immediately before transplant is associated with post-transplant patient survival.¹³ In Park and colleagues, hepatic artery velocity was significantly correlated with MELD score.¹⁴ In Tuladhar and colleagues, in which mean MELD score was 19.28 ± 6.09, no signi-ficant correlations were found between MELD scores and HARI.15 In Baz and colleagues, a significant correlation was found between HARI and MELD scores.¹⁶
The univariate repeated-measures analysis of variance has shown restrictive assumptions about the covariance structure for repeated measures on the same individual. The assumed compound symmetry form for the covariance is inappropriate for longitu-dinal data for at least 2 reasons. First, the constraint on correlation among repeated measurements is somewhat unappealing for longitudinal data, where correlations are expected to decay with increasing separation in time. Second, the assumption of constant variance across time is often unrealistic.¹⁷ The genera-lized estimation equation (GEE) approach is based on a particular type of extension of gene-ralized linear models, known as marginal models, for longitudinal or, more generally, cluster-correlated data.¹⁸
This study aimed to assess the pattern of changes of HARI posttransplant and to identify the role of contributing factors in these patterns. In addition, the study aimed to determine patient risk groups for early arterial thrombosis.

Materials and Methods

Study design
This retrospective study analyzed data from 323 LT recipients, obtained from the Montaserieh Organ Transplant Center registry in Mashhad, Iran, between 2014 and 2021. This research received approval from the Ethics Committee of Mashhad University of Medical Sciences under the Code of Ethics IR.MUMS.REC.1403.289, and all participants gave written informed consent.

Participants
We included all patients aged >18 years who were undergoing deceased donor LT and who were admitted to the postoperative intensive care unit.
Exclusion criteria included patients with acute liver disease, known liver neoplasms, focal liver mass, metastatic liver disease, and previous hepatic surgery. Patients who received a donor liver from a donor younger than 18 years were also excluded from the study.

Data descriptions
Participant characteristics included age, sex, pre-transplant MELD score, diabetes status, history of hypertension, and substance use (smoking and alcohol). Clinical assessments included the hemato-logical markers alanine aminotransferase (ALT), aspartate aminotransferase (AST), and international normalized ratio (INR); in addition, HARI measu-rements were obtained pretransplant and repeated 3 days posttransplant. The MELD score was calculated with the following formula¹⁹: 3.78 × ln(serum total bilirubin [mg/dL]) + 11.2 × ln(INR) + 9.57 × ln(serum creatinine [mg/dL]) + 6.43.

Hepatic artery resistive index measurement
Duplex Doppler sonography was used to study blood flow at level of right and left branches of the hepatic artery, soon after their liver entry, with HARI value obtained as the average of the 2 measurements. HARI was calculated using the following formula: HARI = (peak systolic velocity - end diastolic velocity)/peak systolic velocity.
Patients were prioritized for LT according to disease severity and ranked by MELD score. The transplant procedure was performed in a manner to minimize ischemia time.

Statistical analyses
We presented data as means ± SD for continuous variables and as frequencies (percent) for categorical variables. We used the χ² test to assess associations between categorical variables. We used the Kolmogorov-Smirnov test to examine the normality of the data and the Mann-Whitney U test to determine differences between 2 independent groups. Smooth trajectories were used to visualize patterns of change in HARI over time between the 2 age groups (≤50 years vs >50 years). We used longitudinal analyses to examine the relationship between a response variable and explanatory variables, often including time. The outcome variable for this analysis was a continuous measure of blood flow in the hepatic artery at each occasion (at baseline and repeated measures at 3 days post-LT). Covariates with P < .2 in the χ² and Mann-Whitney tests were included in the regression model. Nonsignificant variables were excluded using backward elimination from the model. The GEE approach was used to estimate the effects of covariates on HARI changes and to determine whether the mean changes in HARI over time differed between age groups. P < .05 was accepted as statistically significant. We analyzed data with SPSS version 27 and R-Studio (4.5.1).

Results

Data from 323 LT recipients were analyzed. Mean age of LT recipients in our study was 48.71 ± 1.022 years. Of these, 216 (66.9%) were male LT recipients, and 260 (80.5%) had diabetes mellitus. The most frequent causes of end-stage liver disease were viral hepatitis in 118 patients (36.5%) and cryptogenic cirrhosis in 74 patients (22.9%). Demographic and clinical para-meters of LT recipients are listed in Table 1.
Hepatic artery resistive index was compared between the 2 LT recipient groups (aged ≤50 years vs aged >50 years) at baseline (pretransplant) and during the first 3 days posttransplant (Figure 1). From Day 2 onward, the groups exhibited different adaptive responses to the transplanted graft, with the greatest difference observed on day 3 after LT.
Trends in mean HARI over time for the 2 age groups showed distinct temporal patterns (Figure 2). Mean HARI increased sharply after LT. In patients aged ≤50 years, HARI rose steeply on day 1 post-transplant. In patients aged >50 years, the increase was particularly pronounced between day 1 and day 3, reaching a peak HARI of 0.835 by day 3, which was a level 0.056 units higher than shown in the ≤50-year age group.
Table 2 shows results of correlation structure matrix of HARI over time. In the longitudinal analyses, to account for the within-subject association among the responses, we considered the first-order autoregressive correlation structure. Regression analy-sis using the GEE approach, with adjusting for potential confounders with P < 0.2, including age, reason for LT, diabetes status, history of hypertension, MELD score, and the clinical markers ALT (day 1 to day 3), AST (day 1 to day), and INR (day 1 to day 3), showed 38.487 as the corrected quasi-likelihood under independence model criterion. After removing nonsignificant confounders, we found that the goodness of fit index decreased to 28.63, indicating a better model fit to data.
Table 3 lists results of the longitudinal analysis using the GEE approach, showing the significant variables and their effects on mean HARI. Compared with deceased donors aged <15 years, recipients of donors aged 51 to 60 years had an average HARI that was 0.074 units higher, whereas recipients of donors aged >60 years had an average HARI that was 0.083 units higher. Levels of AST and ALT showed a negative association with HARI. For each 1-unit increase in ALT and AST, the mean HARI decreased by about 0.0002 units.
Table 4 lists the pairwise comparisons of mean HARI within and between groups. Among patients aged >50 years, the mean HARI was significantly hig-her on day 2 (0.038 units; P = .012) and on day 3 (0.047 units; P = .005) compared with patients aged ≤50 years.

Discussion

This study aimed to explore the patterns of change in HARI and to identify factors influencing these patterns. Our findings showed that HARI tended to increase from pretransplant to 3 days posttransplant across different age groups.
In a study from García-Criado and colleagues, 41 of 90 patients (45.6%) showed a high resistive index in the hepatic artery during the first 72 hours after transplant. Older donor age was found to have a significant effect on the occurrence of a high resistive index.¹⁹ Our study indicated that recipient age may be a factor predisposing patients to high HARI. Distinct temporal patterns emerged among 2 age groups, with the mean HARI showing a steep increase after transplant. In the study from Gaspari and colleagues, recipient age significantly predicted the occurrence of high HARI.⁴ The reason why older recipients are more prone to an early increase in HARI has not been reported elsewhere. In general, older candidates have more comorbidities and long-standing conditions, which may predispose them to more pronounced graft edema and increased vasospasm. Our study included 176 patients (54.5%) who were aged >50 years. Among these older mellitus, 53.9% had primary hypertension, and 60% had secondary hypertension.
Our study found no significant association between pretransplant MELD score and the pattern of change in HARI. Liver disease can alter the blood flow resistance in the hepatic artery, causing it to be either abnormally high or low. Resistance is believed to increase from factors such as swelling from inflammation, physical pressure from regenerative nodules, or stiff liver tissue. Conversely, resistance can decrease due to mechanisms like the “hepatic arterial buffer response” and the formation of abnormal connections between arteries and veins. In a study from Baz and colleagues of 100 patients with cirrhosis, no significant correlation was shown between HARI and MELD scores.¹⁶ In the study by Park and colleagues,¹⁴ HARI did not correlate with MELD score. In contrast, Topal and colleagues²⁰ reported a significant correlation between MELD score and HARI (P < .001, r = 0.616).
Our study revealed an inverse significant asso-ciation between liver function tests (ALT, AST) and HARI, in contrast to the findings from Mousa and colleagues.²¹ The small but significant effects observed for ALT and AST should be interpreted in the context of the GEE framework, which estimates population-averaged associations rather than subject-specific effects. Hepatic artery resistive index reflects intra-hepatic vascular resistance and disease severity, whereas ALT and AST are markers of hepatocellular injury that exhibit substantial intra- and inter-individual variability and may fluctuate indepen-dently of structural or hemodynamic liver changes. Consequently, their independent population-level effects are expected to be modest after adjustment for markers of liver function and disease severity. The modest effects observed for ALT and AST are clinically consistent with their role as markers of hepato-cellular injury rather than fibrosis or hemodynamic impairment. Elevated transaminases may occur in both mild and severe liver disease, whereas HARI is more strongly influenced by intrahepatic vascular resistance and portal hypertension.
In a study from Tian and colleagues in patients with chronic hepatitis B virus, Doppler measurements (eg, portal vein flow/velocities) were altered with fibrosis and inflammation regardless of ALT values, indicating that Doppler parameters may not directly track liver enzyme levels.²² This relationship is complex and paradoxical. However, HARI evaluation may be precise for detection of significant fibrosis in patients with higher liver enzyme levels.²³
Previous studies have investigated the behavior of HARI after LT and the transient increase in HARI has been ascribed to various conditions. Sanyal and colleagues reported that most HARI changes were found to revert to normal within the first postoperative week; deterioration atypical of transient changes requires further evaluation.²⁴
This is the first study to investigate the pattern of changes in HARI across different age groups. Nevertheless, the study had some limitations. Certain donor information, such as ischemia time, was not available, preventing us from assessing the rela-tionship between some donor characteristics and HARI trends.

Conclusion

Our primary finding was the identification of patterns of change in HARI across transplant recipients aged ≤50 years versus >50 years. High-resistance flow in older candidates shortly after LT was commonly observed. Additional research is needed to explore the relationship between donor characteristics and patterns of changes in HARI. An analysis of trends can allow the identification of groups at high risk of arterial thrombosis, to take additional preventive measures in these patients, and to improve the immediate results of treatment in this group of patients. The study highlighted the importance of donor age and recipient age in the HARI changes posttransplant. Without timely intervention, early arterial among older transplant recipients.

References:

1. Capra RP, Lazzarotto-da-Silva G, Grezzana-Filho TJ, et al. Low hepatic artery resistive index on Doppler ultrasound performed on the first post-liver transplant day is associated both with hepatic artery thrombosis and decreased graft survival. Langenbecks Arch Surg. 2023;408(1):231. doi:10.1007/s00423-023-02971-9
   CrossRef - PubMed
2. Hashem M, Medhat MA, Abdeltawab D, Makhlouf NA. Expanding the liver donor pool worldwide with hepatitis C infected livers, is it the time? World J Transplant. 2024;14(2):90382. doi:10.5500/wjt.v14.i2.90382
   CrossRef - PubMed
3. Bekker J, Ploem S, De Jong K. Early hepatic artery thrombosis after liver transplantation: a systematic review of the incidence, outcome and risk factors. Am J Transplant. 2009;9(4):746-757. doi:10.1111/j.1600-6143.2008.02541.x
   CrossRef - PubMed
4. Gaspari R, Teofili L, Mignani V, et al. Duplex Doppler evidence of high hepatic artery resistive index after liver transplantation: role of portal hypertension and clinical impact. Dig Liver Dis. 2020;52(3):301-307. doi:10.1016/j.dld.2019.10.017
   CrossRef - PubMed
5. Piardi T, Lhuaire M, Bruno O, et al. Vascular complications following liver transplantation: a literature review of advances in 2015. World J Hepatol. 2016;8(1):36-57. doi:10.4254/wjh.v8.i1.36
   CrossRef - PubMed
6. Shabunin A, Drozdov P, Levina O, Makeev D, Zhuravel O, Astapovich EY. Negative dynamics of the hepatic artery resistive index as a predictor of early arterial thrombosis after deceased donor liver transplantation. Russian J Transplant. 2022;14(1):10-19.
   CrossRef - PubMed:
   
7. Nickel KJ, Morzycki A, Visser L, Bell E, Ladak A. Effect of magnification in pediatric liver transplantation: a systematic review and meta-analysis. Pediatr Transplant. 2022;26(3):e14223. doi:10.1111/petr.14223
   CrossRef - PubMed
8. Zhao N-B, Chen Y, Xia R, Tang J-B, Zhao D. Prognostic value of ultrasound in early arterial complications post liver transplant. World J Gastrointest Surg. 2024;16(1):13-20. doi:10.4240/wjgs.v16.i1.13
   CrossRef - PubMed
9. Teegen EM, Denecke T, Eisele R, Lojewski C, Neuhaus P, Chopra SS. Clinical application of modern ultrasound techniques after liver transplantation. Acta Radiol. 2016;57(10):1161-1170. doi:10.1177/0284185116633910
   CrossRef - PubMed
10. Kok T, Haagsma EB, Klompmaker IJ, et al. Doppler ultrasound of the hepatic artery and vein performed daily in the first two weeks after orthotopic liver transplantation: useful for the diagnosis of acute rejection? Invest Radiol. 1996;31(3):173-179. doi:10.1097/00004424-199603000-00008
    CrossRef - PubMed
11. Ulu Öztürk F, Tezcan Ş, Ayvazoğlu Soy EH, Uslu N, Haberal M. Effect of meal intake for evaluating hepatic artery by Doppler ultrasonography in liver transplants: does fasting matter for screening hepatic artery due to hemodynamic changes in splanchnic circulation? Clin Transplant. 2022;36(7):e14674. doi:10.1111/ctr.14674
    CrossRef - PubMed
12. Harihara Y, Makuuchi M, Takayama T, et al. Arterial waveforms on Doppler ultrasonography predicting or supporting hepatic arterial thrombosis in liver transplantation. Transplant Proc. 1998;30(7):3188-3189. doi:10.1016/s0041-1345(98)00988-9
    CrossRef - PubMed
13. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology. 2001;33(2):464-470.
    CrossRef - PubMed
14. Park HS, Desser TS, Jeffrey RB, Kamaya A. Doppler ultrasound in liver cirrhosis: correlation of hepatic artery and portal vein measurements with Model for End-Stage Liver Disease score. J Ultrasound Med. 2017;36(4):725-730. doi:10.7863/ultra.16.03107
    CrossRef - PubMed
15. Tuladhar S, Katwal S, Gurung G, Khanal U. Hepatic arterial hemodynamics and model for end-stage liver disease (MELD) scores in chronic liver disease: insights from Doppler ultrasonography. Egypt J Radiol Nucl Med. 2024;55(1):65. doi:10.1186/s43055-024-01220-3
    CrossRef - PubMed:
    
16. Baz AA, Mohamed RM, El-kaffas KH. Doppler ultrasound in liver cirrhosis: correlation of hepatic artery and portal vein measurements with model for end-stage liver disease score in Egypt. Egypt J Radiol Nucl Med. 2020;51(1):228. doi:10.1186/s43055-020-00344-6
    CrossRef - PubMed:
    
17. Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. Applied Linear Statistical Models. 4th ed. McGraw-Hill/Irwin; 1996.
    CrossRef:
    PubMed:
    
18. Hedeker D, Gibbons RD. Longitudinal Data Analysis. John Wiley & Sons; 2006.
    CrossRef:
    PubMed:
    
19. García-Criado A, Gilabert R, Salmerón JM, et al. Significance of and contributing factors for a high resistive index on Doppler sonography of the hepatic artery immediately after surgery: prognostic implications for liver transplant recipients. AJR Am J Roentgenol. 2003;181(3):831-838. doi:10.2214/ajr.181.3.1810831
    CrossRef - PubMed
20. Topal N, Sarkut P, Dündar H, et al. The correlation between Doppler US measurement of hepatic arterial flow and the MELD score in patients with chronic liver disease. Eur Rev Med Pharmacol Sci. 2016;20(2):291-296.
    CrossRef:
    PubMed
21. Mousa N, EL-Eraky A, Arafa M, et al. Value of hepatic artery resistive index in evaluation of liver fibrosis related to non-alcoholic fatty liver diseases. Med J Viral Hepat. 2023;7(2):1-8. doi:10.21608/mjvh.2023.296888
    CrossRef - PubMed:
    
22. Tian L, Tang S, Wang N, Deng H, Zhang Q, Shi T. Hepatic and portal vein Doppler ultrasounds in assessing liver inflammation and fibrosis in chronic HBV infection with a normal ALT level. Front Med (Lausanne). 2023;10:1178944. doi:10.3389/fmed.2023.1178944.
    CrossRef - PubMed
23. Zeng X, Xu C, He D, et al. Influence of hepatic inflammation on FibroScan findings in diagnosing fibrosis in patients with chronic hepatitis B. Ultrasound Med Biol. 2015;41(6):1538-1544. doi:10.1016/j.ultrasmedbio.2015.01.011
    CrossRef - PubMed
24. Sanyal R, Lall CG, Lamba R, et al. Orthotopic liver transplantation: reversible Doppler US findings in the immediate postoperative period. Radiographics. 2012;32(1):199-211. doi:10.1148/rg.321115006
    CrossRef - PubMed