Monitoring peripheral blood CD4+ adenosine triphosphate activity in a liver transplant cohort: Insight into the interplay between hepatitis C virus infection and cellular immunity



Peripheral blood CD4+ adenosine triphosphate [ATP (ng/mL)] release [ImmuKnow Immune Cell Function Assay (ATP)] correlates to immunoreactivity. We hypothesized that ATP levels could provide insight into hepatitis C virus (HCV) infection and recurrent liver disease in liver transplantation (LT). We studied our center's LT cohort, in which ATP levels had been measured off protocol from February 2005 through July 2006. Of the 280 LTs performed since 1993, 114 (58.2%) fit the selection criteria, with a mean age of 49 ± 10 years. LT (alone/combination) indications included HCV (58%), alcoholic liver disease (41%), hepatocellular carcinoma (16%), and other (33%). Four hundred seventy-seven ATP levels were obtained: 3 (1-17) per patient 25 months (4 days to 19 years) from the time from transplantation. Final diagnoses were normal allograft function (n = 166, 35%), recurrent disease (n = 199, 42%), septic event (n = 34, 7%), other (n = 51, 11%), and undetermined (n = 27, 6%). Two hundred eighty-one ATP levels were obtained [3 (1-18) per patient] in 66 HCV(+) patients. Forty-five (68%) developed biopsy-proven recurrent liver disease [188/281 (67%) ATP levels]. The median ATP level (ng/mL) was 162 (1-761); it was lower in HCV(+) patients (151 ± 109) versus HCV(−) patients (211 ± 139; P < 0.0001). ATP ranges in HCV(+) patients were stable from the time from transplantation. In HCV(−) patients, ATP ranges were initially high and eventually decreased to HCV(+) levels (P = 0.01). Immunosuppressant levels were low in 62% of HCV(−) patients versus 38% of HCV(+) patients (P = 0.04). In HCV(+) patients, ATP was lower in disease recurrence (139 ± 97) versus none (181 ± 141; P = 0.01) with similar immunosuppression, and ATP decreased with grade (P = 0.05) but not stage. Time from transplantation, aspartate aminotransferase/alanine aminotransferase >1, and low ATP were independently associated with recurrent HCV. In conclusion, after LT, global cellular immune function appears depressed at baseline in HCV(+) patients versus HCV(−) patients and more so in HCV(+) recurrent disease. Liver Transpl 14:1313–1322, 2008. © 2008 AASLD.

Following liver transplantation (LT), pharmacological suppression of cellular immunity is tailored to achieve optimal graft function with minimal adverse events. Although initial doses of immunosuppressant drugs are high, poorly understood mechanisms of immunotolerance and chimerism occur over time and allow a decrease in dose while graft function is preserved. Preserving graft function comes at a price because immunosuppression increases the risk for opportunistic infections, malignancies, and multiple undesirable effects, including renal toxicity, hypertension, neurotoxicity, diabetes, osteoporosis, and obesity. In addition, these patients are also at risk for either recurrent or de novo liver disease.

Hepatitis C virus (HCV) cirrhosis is the most frequent indication for LT in the United States.1 It also has one of the worst LT survival rates in comparison with other common indications. Recurrence of HCV leads to cirrhosis in 30% of patients within 5 years following LT2; 41% have severe fibrosis or cirrhosis after 6 years.3 Optimal immunosuppression is particularly delicate in patients transplanted for HCV, for whom strong suppression of cell-mediated immunity (CMI) is believed to spur the recurrence of HCV-induced liver disease.4–7

Monitoring CMI is valuable in a variety of applications, including transplantation and management of infectious diseases, autoimmunity, and cancer.8In vitro methods for assessing CMI include the measurement of cell activation signals, lymphoproliferation, cytotoxicity, and cytokine production. Exposure of T-lymphocytes in peripheral blood to a stimulus results in activation and expansion of the T-lymphocytes reactive to that stimulus. Stimulated T-lymphocytes first undergo an influx of ions and increased adenosine triphosphate (ATP) synthesis, which is followed by surface receptor clustering, RNA synthesis, cytokine production and release, and DNA replication.9 In 2002, the US Food and Drug Administration approved a global immune cell function assay (ImmuKnow).10 This assay detects CMI by measuring the concentration of ATP (ng/mL) released from CD4+ cells. Because the CD4+ lymphocytes orchestrate cell-mediated responses through immunoregulatory signaling, the measurement of CD4+ activation is felt to reflect the degree of immune cell function. This test is a functional assay that is reportedly independent of lymphocytes, CD4+ levels, or pharmacokinetics.11, 12

Currently, there are limited data on the use of ImmuKnow in the adult liver transplant population. A recently published multicenter trial,13 including 107 LT reports on correlations between ImmuKnow levels and allograft status, reaffirmed previously described “immune response zones”:11 strong = >524 ng/mL, moderate = 226 to 524 ng/mL, and low = <226 ng/mL. To date, there are no reports on ImmuKnow dynamics in populations transplanted for HCV. There is some evidence, however, that HCV infection may depress CD4+ response14, 15 and that this is associated with failure of treatment16, 17 and progression to cirrhosis.18

ImmuKnow has been available at the Loma Linda University Medical Center (LLUMC) since November 2004 for renal transplant patients and since February 2005 for liver transplant patients. In this article, we present the results of a review of our experience with this assay in our LT population, with an emphasis on associations with HCV infection and recurrent liver disease. We hypothesized that recurrent HCV would be associated with a low immune response zone.


A1ATD, alpha-1 antitrypsin deficiency; ACR, acute cellular rejection; AIH, autoimmune hepatitis; ALD, alcoholic liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATP, adenosine triphosphate; BDO, bile duct obstruction; CCR, chronic cellular rejection; CI, confidence interval; CMI, cell-mediated immunity; Crypt, cryptogenic; CSA, cyclosporine; DT, drug toxicity; FH, fulminant hepatitis; FK, tacrolimus; HAT, hepatic artery thrombosis; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LLUMC, Loma Linda University Medical Center; LT, liver transplantation; NAF, normal allograft function; OR, odds ratio; ORLD, obesity-related liver disease; PBC, primary biliary cirrhosis; PSC, primary biliary cirrhosis; RD, recurrent disease; SD, standard deviation; SE, septic event; ULN, upper limit of normal; Undt, undetermined; WBC, white blood cell; WD, Wilson's disease.


We retrospectively reviewed our liver transplant cohort in which ImmuKnow had been performed on an off-protocol basis since February 2005. This study was approved by the LLUMC institutional review board (OSR #56080).


Patients were included if they were followed after LT in our center and had one or more ImmuKnow levels obtained until July 2006. Patients were identified by the screening of our laboratory's data set for ImmuKnow levels. Patients younger than 18 years old at the time of ImmuKnow testing or who had had combined liver and kidney transplantations were excluded.

Collected Data

Data were compiled into an electronic relational database (FileMaker Pro 8 Advanced) designed specifically for the study. Baseline data pertinent to the time of LT included the date of LT, demographics, cause of liver disease, immunosuppression regimen, and HCV quantitative polymerase chain reaction when applicable. For every ImmuKnow level obtained, data related to the time of testing included the time from LT, indication for ImmuKnow testing, immunosuppression regimen, physical examination, routine biochemistry, HCV quantitative RNA levels when available, and final diagnosis. All ascertainable data were considered attributable to the date of ImmuKnow testing if they had been obtained within 2 weeks (up to 7 days before or after the ImmuKnow level). In the final analyses, HCV quantitative RNA levels were not studied because of the disparity in assays used over the study period and differences in units.


Each ImmuKnow level was termed an “event”; an “event database” was created in which data were collected relevant to each event.

The identifiable indications for ImmuKnow testing were based on usual clinical indices of suspicion. “Abnormal liver enzymes” referred to events in which previously normal enzymes [≤1 × upper limit of normal (ULN)] had become elevated (>1 × ULN), without any sign for a possible diagnosis. “Liver enzymes above baseline” referred to events in which liver enzymes had risen significantly (≥2-fold) above an abnormal baseline (for example, in patients with recurrent HCV). Events in patients with normal liver enzymes and function who had had intentionally repeated testing (although without any institutional protocol) were labeled as “normal allograft function.” However, events for which there had been no clearly identifiable indication were labeled as “not specified.” Tacrolimus (FK) and cyclosporine (CSA) trough levels were defined for the purposes of the study as “optimal” (FK: 5-9 ng/mL; CSA: 125-275 ng/mL), “low” (FK: < 5 ng/mL; CSA: < 125 ng/mL), or “high” (FK: > 9 ng/mL; CSA: > 275 ng/mL).

The final diagnoses associated with each ImmuKnow level were called “event diagnoses.” Acute cellular rejection was defined by the 9-point Banff rejection activity index19 as mild, moderate, or severe. All cases diagnosed as recurrent HCV referred to those with HCV-related liver disease identified on a liver biopsy either previous to or at the time of the ImmuKnow testing. Straightforward cases of recurrent HCV were described by the METAVIR system.20 The sometimes difficult distinction between histologically ambiguous cases of recurrent HCV and acute cellular rejection had been made by the associated clinical presentation, with rectification of the diagnosis during follow-up as often is the case in clinical practice. Correlations between ImmuKnow levels and histology were made only in cases in which a liver biopsy had been performed within 7 days prior to or following an ImmuKnow level.

The diagnosis of “septic event” was based on clinical evidence of infection or sepsis, whether by an opportunistic agent or not [for example, fever, viral syndrome, leukocytosis, and positive blood cultures (bacterial, viral, or fungal)], followed by an appropriate response to treatment. The diagnoses of chronic rejection, bile duct obstruction, de novo liver disease, and hepatic artery thrombosis were based on standard clinical, histological, and/or imaging criteria. “Drug toxicity” referred to adverse events attributed to immunosuppressant drugs. Event diagnoses that still remained equivocal at the time of data collection were re-reviewed, and either a consensus was achieved or they were considered “undetermined.”

Quality Control

All included patients had at least one if not all of their ImmuKnow levels performed at the LLUMC clinical laboratory. Some were performed in outside laboratories. All laboratories performing the assay had been licensed by Cylex. The assay was performed on sodium heparin anticoagulated whole blood samples within 30 hours of blood draw, as previously described.13 Data were collected by chart review and were cross-referenced with the institution's electronic medical result documentation systems (biochemistries, clinic visits, hospitalizations, pathology, and imaging). Data were collected onto hard-copy forms and then entered into the electronic database by 1 person. All liver biopsies had been previously evaluated by 1 experienced liver pathologist. Liver biopsies with equivocal results were re-examined, and the examiner was blinded to clinical data.


Data were treated with Statview and JMP software from SAS. Quantitative data were described as the mean ± standard deviation or mean ± 95% confidence interval (CI) for normal distributions, the median (extremes) for nonnormal distributions, or both when appropriate. Univariate analyses were performed by chi-square, t, or nonparametric (Mann-Whitney and Kruskal-Wallis) tests as appropriate. Multivariate analyses are described later.


Patient Cohort

From 1993 to July 2006, 280 LTs were performed at LLUMC. At the time of this study, 196 (70%) LT patients of adult age were being followed, including patients transplanted elsewhere. Of these, 114 (58.2%) fit our selection criteria. Patient demographics at the time of LT and indications are described in Tables 1 and 2. There was a significant predominance (P = 0.0547) of middle-aged Hispanic males, mostly of Mexican American origin. The most prevalent indication for LT was HCV: 28 patients (28.6%) presented exclusively with HCV, and 66 (57.9%) had chronic HCV with or without another cause of liver disease. Only 11 (9.6%) presented exclusively with alcoholic liver disease (ALD); however, 47 (41.2%) had ALD with or without another cause of liver disease. Eighteen patients (15.8%) had hepatocellular carcinoma.

Table 1. Patient Demographics of the Entire Study Group
Variablen (%)Males: n (%)Age: Mean ± SD
  1. Abbreviation: SD, standard deviation.

Total114 (100)79 (69.3)49.2 ± 10.4 years
African American4 (3.5)2 (50.0)43.3 ± 17.4 years
American Indian1 (0.9)1 (100)49.0 years
Asian5 (4.4)3 (60.0)50.6 ± 4.6 years
Caucasian42 (36.8)23 (54.8)47.3 ± 13.1 years
Hispanic62 (54.4)50 (80.6)50.5 ± 7.9 years
Table 2. Indications for Liver Transplantation in the 114 Recipients
Liver Diseasen (%)Liver Diseasen (%)
  1. Abbreviations: A1ATD, alpha-1 antitrypsin deficiency; AIH, autoimmune hepatitis; ALD, alcoholic liver disease; Crypt, cryptogenic; FH, fulminant hepatitis; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; ORLD, obesity-related liver disease; PBC, primary biliary cirrhosis; PSC, primary biliary cirrhosis; WD, Wilson's disease.

HCV28 (28.6)HCV/HBV2 (1.8)
HCV/ALD24 (21.1)WD2 (1.8)
ALD11 (9.6)Crypt2 (1.8)
HCV/HCC9 (7.9)ALD/HCV/HBV2 (1.8)
HBV5 (4.4)HCC1 (0.9)
ALD/HCC4 (3.5)Crypt/HCC1 (0.9)
AIH4 (3.5)HCV/ALD/ORLD1 (0.9)
PBC4 (3.5)ALD/ORLD/HCC1 (0.9)
ALD/ORLD3 (2.6)PBC/HCC1 (0.9)
PSC3 (2.6)A1ATD1 (0.9)
FH3 (2.6)ALD/A1ATD/HCC1 (0.9)
  Biliary Atresia1 (0.9)

Indications for Testing

From these 114 patients, a total of 477 ImmuKnow levels were obtained: 3 (1-17) per patient at a median of 25 months (4 days to 19 years) of follow-up from LT. The most frequent reasons for ordering the ImmuKnow levels were assessment of recurrent liver disease [177/477 (36.8%)] and management of patients with normal allograft function [141/477 (29.6%)] The remaining indications were sepsis (8.4%), unspecified (7.6%), abnormal liver enzymes (5.7%), liver enzymes above baseline (3.6%), chronic rejection (2.7%), bile duct obstruction (2.5%), drug toxicity (1.5%), acute cellular rejection (1.3%), de novo liver disease (0.2%), and hepatic artery thrombosis (0.2%). At the time of these 477 tests, the principle immunosuppression was FK (n = 357, 74.8%), CSA (n = 78, 16.4%), sirolimus (n = 33, 6.9%), microphenolate mophetil (n = 3, 0.6%), or none (n = 6, 1.3%).

Final Diagnoses

Figure 1 demonstrates that the most frequent final event diagnoses were recurrent liver disease [199/477 (41.7%)] and normal allograft function [166/477 (34.8%)]. HCV was responsible for 188 of 199 (94%) events of exclusive recurrent disease; the remaining 11 were ALD (4) in 1 patient, primary biliary cirrhosis (4) in 2 patients, and hepatitis B virus (3) in 1 patient. There were relatively few diagnoses of acute or chronic rejection, bile duct obstruction, or hepatic artery thrombosis. There were only 34 (7.1%) septic events. Twenty-seven events (5.7%) remained without a clearly determined diagnosis.

Figure 1.

Mean ± 95% CI ImmuKnow levels [ATP (ng/mL)] according to 477 event diagnoses in 114 liver transplantation patients. The classification of diagnoses [n (%), mean/standard deviation/median/extremes] was as follows: ACR [10 (2.1), 151/79/163/23-237], BDO [15 (3.1), 145/107/110/7-427], CCR [13 (2.7), 131/134/74/13-437], DT [12 (2.5), 268/158/210/76-568], HAT [1 (0.2), 55], NAF [166 (34.8), 214/140/191/2-761], RD [199 (41.7), 143/97/137/1-380], SE [34 (7.1), 159/111/172/11-396], and Undt [27 (5.7), 216/163/181/2-644]. The significance of differences (for P < 0.05) is shown in brackets (unpaired t test). The horizontal dashed line separate the “low immune response zone” (<226 ng/mL ATP) from the “moderate immune response zone” (226-524 ng/mL ATP) as previously described.13Abbreviations: ACR, acute cellular rejection; ATP, adenosine triphosphate; BDO, bile duct obstruction; CCR, chronic cellular rejection; CI, confidence interval; DT, drug toxicity; HAT, hepatic artery thrombosis; NAF, normal allograft function; RD, recurrent disease; SE, septic event; Undt, undetermined.

Of the 66 patients who had HCV at the time of LT (alone or in combination with other causes of liver disease), 45 (68.2%) had developed some degree of HCV recurrent disease following LT. Of the 281 events occurring in these 66 patients, 188 (66.9%) were attributable to recurrent disease. Patient demographics according to HCV status at the time of LT and following LT are shown in Table 3. There were no significant differences in terms of gender and race. The subgroup of 66 HCV(+) patients had a median follow-up of 25 months (9 days to 12 years) from the time of LT. Recipient characteristics at the time of LT were as follows: 50.2 ± 7.7 years of age, 78.8% male, 60.6% Hispanic, 39.4% associated ALD, 13.6% hepatocellular carcinoma, and a median HCV viral load of 631 × 103 UI/mL (0 to >28,800 × 103).

Table 3. Patient Demographics and ImmuKnow Levels According to HCV Status at LT and Following LT
VariableIndication for LT: HCV(−)Indication for LT: HCV(+)Post-LT Histological HCV Recurrence: NoPost-LT Histological HCV Recurrence: Yes
  • Abbreviations: ATP, adenosine triphosphate; HCV, hepatitis C virus; LT, liver transplantation.

  • *

    Mean/standard deviation/median/extremes.

  • versus

  • P < 0.0001.

  • §

    versus ∥P = 0.01 (Mann-Whitney).

n (%)48 (42.1)66 (57.9)21/66 (31.8)45/66 (68.2)
Age (years) at LT*47.9/13.2/51.5/4–6550.2/7.7/49.5/25–7150.2/7.6/52.0/36–6450.1/7.9/49.0/25–71
Males [n (%)]27 (56.3)52 (78.8)16/21 (76.2)36/45 (80.0)
African American2 (4.2)2 (3.0)1 (4.8)1 (2.2)
American Indian01 (1.5)1 (4.8)1 (2.2)
Asian3 (6.3)2 (3.0)02 (4.4)
Caucasian21 (43.8)21 (31.8)6 (28.6)15 (33.3)
Hispanic22 (45.8)40 (60.6)14 (66.7)26 (57.8)
ImmuKnow [n (%)]19628163/281 (22.4)218/281 (77.6)
ImmuKnow*[ATP (ng/mL)]211/140/195/2–761151/109/139/1–563182/127/167/7–557§141/103/134/1–563∥

ImmuKnow Levels: Descriptive and Univariate Analyses

The mean ImmuKnow level was 176 ± 126 ng/mL ATP. ImmuKnow levels according to event diagnoses are shown in Fig. 1 and Table 3. Of note, levels in recurrent disease were significantly lower than those of normal allograft function or drug toxicity. On the whole, mean ImmuKnow levels in our population were in the previously described “low immune response zone”10 (<226 ng/mL ATP).

Of note, ImmuKnow levels were remarkably low in events of acute and chronic rejection and bile duct obstruction and high in cases of drug toxicity, but there were relatively few such cases, and this introduced type 1 error. Septic events tended to have levels similar to if not higher than those in patients with normal allograft function. Of these 34 events, 19 had well-documented sources, most of which involved bacterial infections. There were 17 events from 3 patients with posttransplant lymphoproliferative disease. ImmuKnow levels tended to be lower in this small subgroup (134 ± 93.1 ng/mL ATP) compared to those without posttransplant lymphoproliferative disease (177 ± 127 ng/mL ATP), but this did not reach significance, and only 1 case was associated with Epstein-Barr virus infection.

ImmuKnow levels were significantly lower (P < 0.0001) in HCV(+) recipients (151 ± 109 ng/mL ATP) compared to HCV(−) recipients (211 ± 140 ng/mL ATP).

ImmuKnow Levels: Relation to the Time from LT and HCV(+) and HCV(−) Patients

As shown in Fig. 2A, following LT, there was a significant trend of a decrease in ImmuKnow levels over time. Interestingly, Fig. 2B and Table 4 show that this trend was essentially observed in HCV(−) recipients, whereas HCV(+) recipients tended to have stable levels over time and almost consistently significantly lower levels compared to HCV(−) recipients.

Figure 2.

Mean ± 95% CI ImmuKnow levels according to 477 event diagnoses in 114 liver transplantation patients classified by time from liver transplantation (months and years). (A) Grouped results. (B) Results according to HCV(+) or HCV(−) recipients. Table 4 shows descriptive statistics and group differences. Abbreviations: ATP, adenosine triphosphate; CI, confidence interval; HCV, hepatitis C virus.

Table 4. ImmuKnow Levels [Adenosine Triphosphate (ng/mL)] According to the Time from LT and Recipient HCV Status
 Time from LTP
0 to 3 Months3 to 6 Months6 to 12 Months1 to 3 Years>3 Years
  • Abbreviations: HCV, hepatitis C virus; LT, liver transplantation; SD, standard deviation.

  • *

    Student t test.

  • Mann-Whitney test.

  • Kruskal-Wallis test for intragroup differences.

All 114 LT recipients      
 n levels (% of 477)69 (14.5)53 (11.1)43 (9.0)118 (24.7)194 (40.7) 
48 HCV(−) LT recipients      
 n levels (% of 196)25 (12.7)23 (11.7)21 (10.7)51 (26.0)76 (38.8) 
66 HCV(+) LT recipients      
 n levels (% of 281)44 (15.7)30 (10.7)22 (7.8)67 (23.8)118 (42.0) 
P [HCV(−) versus HCV(+)]*0.00310.07920.07200.01700.0033 
P [HCV(−) versus HCV(+)]0.00940.00390.05810.01270.0115 

This strong trend toward low immunoreactivity in HCV(+) recipients was not due to a relative increase in pharmacological immunosuppression because CSA or FK trough levels were low in 62% of HCV(−) recipients versus 38% of HCV(+) recipients (P = 0.04). In 315 events, FK was the primary immunosuppressant, and trough levels (ng/mL) were significantly lower (P = 0.0134) in HCV(+) recipients (n = 215, mean: 6.7 ± 3.0, median: 5.8, extremes: 3.0-18.5) compared to HCV(−) recipients (n = 137, mean: 7.6 ± 3.6, median: 6.9, extremes: 3.0-23.2). As shown in Fig. 3, FK troughs decreased significantly over time from LT for both HCV(+) and HCV(−) recipients (P < 0.0001) and trended toward being lower in HCV(+) recipients at each time interval from LT. Similarly, when both CSA and FK levels were assessed over time, HCV(+) recipients with histological recurrence were more often in the “low trough level” category than HCV(−) recipients (P = 0.0003).

Figure 3.

Levels of pharmacological immunosuppressants at the time of ImmuKnow testing in the 114 recipients, stratified according to the time from liver transplantation and HCV status at transplantation. (A) Mean ± 95% CI FK troughs [HCV(+): n = 215/352; HCV(−): n = 137/352]. (B) Distribution according to CSA or FK troughs [HCV(+): n = 260/435; HCV(−): n = 175/435] in “high,” “optimal,” and “low” ranges (see the Patients and Methods section for definitions). Kruskal-Wallis [P < 0.0001 for HCV(+) and HCV(−)] and chi-square tests (P = 0.0003) were used. Abbreviations: CI, confidence interval; CSA, cyclosporine; FK, tacrolimus; HCV, hepatitis C virus.

Factors Associated with Post-LT Recurrent HCV

Disease recurrence was established from liver biopsies (regardless of enzyme patterns or associated events) performed either before or at the time of testing of ImmuKnow levels. The 45 of 66 patients who developed recurrent HCV liver disease did not differ significantly from the 21 free of recurrent HCV in any demographic variable (Table 3). Organ donor characteristics were not studied.

A total of 281 ImmuKnow levels were obtained: 3 (1-17) per HCV(+) patient. Patients with and without HCV recurrence had a comparable number of events (levels): 218 [3 (1-17)] versus 63 [3 (1-8)].

ImmuKnow levels were significantly lower in disease recurrence (141 ± 103 ng/mL ATP) versus no recurrence (182 ± 127 ng/mL ATP; P = 0.01). As shown in Table 5, compared to recipients without histological recurrence, those with histological recurrence presented significantly later from the time of LT and had a higher body mass index, higher aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels, higher AST/ALT ratios, lower white blood cell counts but higher lymphocyte counts, and lower FK trough levels. No other variables studied were significantly different between the 2 groups.

Table 5. Significant Differences Between HCV(+) Recipients Who Presented (n = 45) or Did Not Present (n = 21) with Histological Recurrence Following LT: 281 Events in 66 Patients
VariablenPost-LT Histological HCV Recurrence: NoPost-LT Histological HCV Recurrence: YesP*
  • NOTE: Data are expressed as n/mean/standard deviation/median/extremes. n represents the number of events with available data for each variable.

  • Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATP, adenosine triphosphate; BMI, body mass index; FK, tacrolimus; LT, liver transplantation; ULN, upper limit of normal; WBC, white blood cell.

  • *

    Mann-Whitney test.

Time from LT (m)28163/6.8/13.3/2.3/0.3–66.1218/43.0/31.5/37.7/1.2–144<0.0001
BMI (kg/m2)21553/24.5/5.4/24.3/19.5–38.9163/27.5/5.4/27.2/16.5–42.50.001
AST (×ULN)27163/1.5/2.1/0.9/0.4–13.4208/2.4/2.2/1.7/0.5–15.3<0.0001
ALT (×ULN)27163/2.0/3.1/1.0/0.3–19.5208/2.0/1.6/1.4/0.2–8.40.025
AST/ALT (×ULN/×ULN)27163/1.0/0.3/1.0/0.2–1.5208/1.4/0.7/1.2/0.4–4.20.0002
WBC (×106/μL)27463/5.9/2.4/5.7/0.8–12.7211/5.0/1.9/4.9/1.0–10.40.0129
Lymphocytes (×106/μL)26559/1.1/0.7/1.0/0.1–4.0206/1.3/0.7/1.2/0.2–3.80.0096
FK trough (ng/mL)21548/8.2/3.5/8.8/4.0–18.5167/6.2/2.7/5.3/3.0–18.20.0002
ImmuKnow [ATP (ng/mL)]28163/182/127/167/7–557218/141/103/134/1–5630.01

Induction and maintenance immunosuppression regimens were not significantly different between the 2 groups. Over the study period, CSA and FK trough levels for HCV(+) recipients with recurrent disease were high in 14.4%, optimal in 36.2%, and low in 39.4% of cases versus 17.2%, 50.0%, and 32.8% in recipients without recurrent disease, respectively (P = 0.07).

Only 44 ImmuKnow levels were drawn in 29 patients within a 7-day period before or after a liver biopsy [median: 4 (−3 to 7 days)]. Ten levels were associated with the diagnosis of acute cellular rejection, 8 of which were in 4 HCV(+) patients. Thirty-four levels corresponded to disease recurrence in 25 HCV(+) patients, and biopsies were performed at a median of 33.6 months from transplant (extremes: 1.2-106 months). In these cases, ImmuKnow levels (132 ± 97 ng/mL ATP, median: 162) were significantly lower (P = 0.05) in cases of METAVIR grade 2 (n = 14, 122 ± 111 ng/mL ATP, extremes: 1-302) versus grade 1 (n = 20, 140 ± 87 ng/mL ATP, extremes: 1-239). There were no cases of grade 3 or 4. Levels did not significantly differ according to METAVIR stages 1 to 4 yet there was a trend (P = 0.076, Cramer's V = 0.506) toward a correlation between grade and stage: grade 1 was associated with stages 0 (47.4%), 1 (31.6%), 2 (10.5%), 3 (5.3%), and 4 (5.3%), whereas grade 2 was associated with stages 0 (7.1%), 1 (42.9%), 2 (35.7%), 3 (14.3%), and 4 (0%).

The remaining 41 HCV(+) patients had nonsynchronous biopsies, all >7 days before ImmuKnow testing, corresponding to 247 levels. In the 27 of 41 (65%) who had recurrence, liver biopsies had been performed at a median of 42.5 months from transplant (extremes: 3.3-116 months). The mean ImmuKnow level in this subgroup (n = 174) was 140 ± 97 ng/mL ATP (extremes: 2-563). This was comparable to levels obtained in cases of HCV recurrence diagnosed with synchronous biopsies.

In this cohort, only 5 patients were being treated for HCV at a median of 19.9 months from transplantation (extremes: 12.8-71.5 months), corresponding to only 10 ImmuKnow levels [60 ± 49 ng/mL ATP (extremes: 3-146)]. These were significantly lower than the remaining 467 levels [178 ± 126 ng/mL ATP (extremes: 1-761), P = 0.0035], the 271 levels corresponding to HCV(+) patients not on treatment [154 ± 109 ng/mL ATP (extremes: 1-563), P = 0.0078], and the 208 levels corresponding to HCV(+) patients with recurrence but not on treatment [146 ± 103 ng/mL ATP (extremes: 1-563), P = 0.0098]. The exclusion of these 10 levels associated with HCV treatment from the cohort of 477 levels did not significantly change any differences or trends reported previously.

As could be expected, the frequency of events related to the diagnosis of HCV recurrence increased with time from LT: 15.9% within 3 months, 56.7% in the 4th to 6th months, 72.7% in the 7th to 12th months, 95.5% in the 13th month to 3rd year, and 96.6% thereafter. This trend can be explained by both an increase in the incidence of recurrent disease and a decrease in other LT complications as follow-up progressed. The suspicion of recurrent HCV dominated the indications for ImmuKnow testing, particularly 12 months after LT. After this point, indications were mainly for suspected recurrence, normal allograft function, or unspecified reasons, almost all of which were associated with a final diagnosis of recurrence.

In multivariate analysis using continuous variables, only years from LT [odds ratio (OR): 3.2, 95% CI: 2.1-4.7, P < 0.0001] and the AST/ALT (×ULN/×ULN) ratio (OR: 4.1, 95% CI: 1.7-9.8, P = 0.0018) were independently and positively associated with HCV histological recurrence. When variables were coded into bimodal nominal variables, 3 years from LT (OR: 15.9, 95% CI: 5.4-47.6, P < 0.0001), an AST/ALT ratio > 1 (OR: 2.1, 95% CI: 1.1-4.3, P < 0.03), and an ImmuKnow level < 125 ng/mL ATP (OR: 3.4, 95% CI: 1.7-6.6, P = 0.0004) were independently associated with HCV histological recurrence.


We describe our experience with the ImmuKnow global immune function assay (Cylex), having measured peripheral blood CD4+ ATP (ng/mL) release in 114 adult LT patients who had a total of 477 tests. Our findings suggest that HCV(+) liver transplant recipients have a significantly lower state of immunoreactivity unrelated to pharmacological immunosuppression, and this may be a risk factor for recurrent liver disease.

First, in contrast to previous reports,11, 13 we did not observe well-demarcated ranges in ImmuKnow levels that correlated with sepsis (suggesting overimmunosuppression) or rejection (suggesting underimmunosuppression). We observed, in fact, a strong trend toward low ImmuKnow levels in the small group of patients with acute or chronic rejection, which is counterintuitive to the principal hypothesis in previous reports. Our results may reflect that, in these cases of rejection, ImmuKnow testing was performed when attempts to reverse the process had already been implemented. Hence, the low levels do not necessarily reflect the immune status leading to the rejection process. The low ImmuKnow levels observed in the cases of bile duct obstruction, although few in number, were significant. To our knowledge, this has not been described before, and its significance is unclear.

Contrary to our expectation, the ImmuKnow levels in cases of sepsis tended to be similar to if not higher than those in cases of normal allograft function, although with a very large range. This may simply reflect the very broad definition of “septic event” used in this study. These discrepancies may also be due to type 1 error.

On the other hand, we observed lower ImmuKnow levels in patients with recurrent liver disease (143 ± 97 ng/mL ATP), most of which were due to HCV (Fig. 1). Current reports on the experience of ImmuKnow and LT have not directly addressed this issue. However, there are reports of significant associations between low levels and viral infections, including cytomegalovirus and Epstein-Barr virus infection, or reactivation in patients with liver, kidney, and pancreas allografts.12, 13, 21, 22

The most significant finding of our study is the persistently low immunoreactivity observed in HCV(+) LT recipients versus HCV(−) recipients. Immunoreactivity in the former was low from the time of transplantation, and this persisted throughout the course of the post-LT follow-up. These lowered ImmuKnow levels were not due to stronger pharmacological immunosuppression at any point in the post-LT course; in fact, HCV(+) recipients tended to have significantly lighter immunosuppression regimens, as is now becoming customary. In contradistinction, HCV(−) recipients, transplanted mostly for alcoholic or autoimmune liver disease, had significantly higher immunoreactivity, especially during the first year following LT. However, with time, their ImmuKnow levels gradually decreased to a level comparable to that of HCV(+) recipients after the third year (Fig. 2).

Furthermore, we found that within the subset of HCV(+) LT recipients, those with recurrent disease had significantly lower ImmuKnow levels than those who remained HCV RNA–positive but without documented recurrent disease. We could not demonstrate a relationship between ImmuKnow levels and HCV RNA levels because of a lack of adequate data. However, in a limited subset of patients with liver biopsies timed closely to ImmuKnow levels, we did observe a significant inverse relationship between these levels and HCV-related histological activity scores, which in turn correlated with higher stages. Hence, we surmise that low immunoreactivity is associated with more intense recurrence of HCV-related liver disease.

In multivariate analysis, increased time from LT, an AST/ALT ratio > 1, and an ImmuKnow level < 125 ng/mL ATP were independently associated with recurrent HCV disease. The increased incidence and progression of post-LT HCV recurrent liver disease with time have been previously reported.2, 3 Similarly, increased AST and, in particular, persistently elevated transaminases have also been previously described as indicators of recurrent disease.3 To our knowledge, our observation that a depressed state of immunoreactivity is independently associated with recurrent disease is the first of this kind.

This retrospective, descriptive study does not explain the mechanisms involved in the apparent lowered immunoreactivity in HCV(+) LT recipients. We hypothesize that perhaps it is a result of interactions between the virus and CD4+ cells, as previously suggested in the nontransplant setting.14, 15 Furthermore, such a relationship has been attributed to treatment failure of HCV.16, 17 Lastly, depression of CD4+ cells in HCV(+) patients has been associated with the progression of chronic hepatitis to cirrhosis18; thus, perhaps a parallel can be drawn with our observation of progressive disease following LT. Taken together, perhaps such interactions allow HCV to replicate free of host response, with immune-mediated hepatitis developing during phases of heightened immunoreactivity.

The implications of our observations are potentially far-reaching. Currently, there is much debate over the role of pharmacological immunosuppression in the progression of recurrent post-LT HCV disease and, in particular, the role of steroids. It would appear that high-dose boluses of corticosteroids used to treat acute cellular rejection might spur HCV replication with subsequent hepatitis during the immune-reactivation phase.5 Within the LT community, there is agreement that steroids should be tapered off more rapidly following LT in HCV(+) patients as opposed to HCV(−) patients.4 However, there are no clear data concerning what dose/time protocol to follow and what criteria should be met; in particular, too rapid steroid tapering may in fact spur an immune-reactivation phase leading to hepatitis.23, 24, 25 The hypothetical benefits of steroid-free post-LT induction and long-term immunosuppression do not include decreased progression of recurrent disease; instead, there is a higher risk of progression due to a higher incidence of acute cellular rejection requiring steroid boluses.26 We suggest that monitoring ImmuKnow levels in HCV(+) patients following LT may afford a better appreciation of innate and pharmacologically immunosuppressed immunity and could help avoid overimmunosuppression and the fluctuations that may spur HCV replication on the one hand and immune-mediated hepatitis on the other hand.

Our study has the limitations of a retrospective study. ImmuKnow levels were drawn at sometimes haphazard times and at different frequencies per patient, and in 7.6% of cases, no clear reason for testing could be identified. This reflects the differences in the practices of the physicians. Despite a careful review of charts, 5.7% of cases had no diagnosis, and this reflected either inadequate ancillary studies or questionable indications for ImmuKnow testing. The small number of allograft rejection episodes coupled with late ImmuKnow testing rendered the analysis of this subgroup very difficult. Perhaps our all-inclusive criteria for sepsis can explain to some degree the disparity of ImmuKnow levels for this diagnosis.

Despite these caveats, we believe that this study provides insight into the understanding of recurrent HCV in liver allografts through the measurement of global immune function testing by ImmuKnow levels. Low levels alone may not always correlate to recurrent HCV disease at the time of testing but may herald the development of recurrence thereafter. Ultimately, such testing could assist in tailoring immunosuppression to avoid HCV recurrence. The potential of this assay should be further studied in large prospective trials, including repeated testing at regular intervals and at index diagnoses. It may prove useful to test patients immediately before transplantation, as their immune status may be a preconditioning phenomenon, impacting the transplant outcome. Lastly, in light of our findings, we would also suggest that ImmuKnow testing should be studied in the setting of nontransplant HCV management, possibly providing insight into the progression of disease and/or response to immune-modulating treatments.


The authors thank Nancy Chang, Pharm.D., for data retrieval; Mrs. Marisela Santiago for electronic database development and data entry; and Miss Elissa Flores for help in retrieving ImmuKnow results.