Objectives: Cytomegalovirus and Epstein-Barr virus remain leading causes of morbidity and mortality in the living-donor liver transplant population, particularly in pediatric patients. Herein we compare the incidence, timing, and risk factors for infection in this group.
Materials and Methods: We performed a retrospective study of 344 consecutive pediatric patients 193 women (56.1%) who received living-donor liver transplants at Kyoto University Hospital. Patients were followed-up for maximum 7.1 ± 3.6 years (range, 0.02-13.2 y) after surgery.
Results: The mean age at the time of transplant was 3.95 ± 4.75 years (median, 1.38 y; range, 0.07-17.87 y). A total of 156 patients (45.2%) developed viral infections. Of those patients, 91 (26.5%) developed cytomegalovirus infection, and 93 (27%) developed Epstein-Barr virus. Cytomegalovirus developed at 39.3 ± 34.6 days, while Epstein-Barr virus developed 3.99 ± 3.67 years after transplant. Frequent rejection attacks (hazard ratio [HR],1.58; 95% confidence interval [CI]: 0.14-2.18; P = .006) were an independent predictor for postoperative cytomegalovirus infection, while preoperative cytomegalovirus seropositive results (HR, 1.76; 95% CI: 1.03-2.18; P = .038), short cold ischemia time (HR, 1.0; 95% CI: 0.99-1.0; P = .02), larger graft (HR, 1.3; 95% CI: 1.00-1.73; P = .047), and new cases compared to old cases (HR, 2.27; 95% CI: 1.14-4.52; P = .019) were independent predictors for postoperative Epstein-Barr virus infection.
Conclusions: Extended surveillance of cytomegalovirus and Epstein-Barr virus DNAemia is recommended for pediatric patients receiving living-donor liver transplants, particularly infants who are at high risk, and especially those exposed to frequent attacks of rejection and those that receive larger grafts.
Key words : Hepatic grafts, Immunosuppression, Rejection, Risk factors
Liver transplants have been successfully used to treat children with end-stage liver disease, offering the opportunity for a long, healthy life.1 Living-donor liver transplants (LDLT) have come to account for a substantial number of pediatric cases performed in many centers throughout the world where deceased-donor organ procurement is rare.2 Immunosuppressive drugs are used to prevent rejection, inhibit activation of T lymphocytes, and modulate cell proliferation and macrophage function, thereby creating an optimal environment for the development of infections. Thus, infectious complications now represent the most common cause of morbidity and mortality after transplant procedures. Early and severe viral infections are caused by viruses of the herpes family, including Epstein-Barr virus (EBV), and cytomegalovirus (CMV).3 Significant scientific breakthroughs and remarkable advances in molecular diagnostics and therapeutics have reduced the incidence and severity of CMV disease and EBV-related posttransplant lymphoproliferative disorders (PTLD) after liver transplant surgery during the early postoperative period. A parallel decline in associated morbidity and mortality has followed. However, despite these improvements, CMV and EBV remain common infectious complications and continue to negatively influence the outcome of liver transplant procedures.4 Moreover, the widespread and prolonged use of antiviral drugs changes the natural course of CMV disease by delaying its onset. In addition, an increased incidence of antiviral drug-resistant CMV infections now exists.5 In contrast, experimental and clinical data demonstrate a promising role for immunotherapy in preventing and treating PTLD and advocate the role of optimal preventive and treatment strategies for EBV to reduce the incidence of PTLD.6-8
In our current study, we highlighted the most important aspects distinguishing CMV and EBV. In addition to the difference in the timing of appearance, the increasing frequency of EBV but not CMV infection over years, possibly due to immune response, is being recognized in the pathogenesis of CMV and EBV infection after liver transplant procedures. Such findings should provide additional avenues and opportunities for improving disease management strategies. Preoperative selection of pediatric patients for LDLT should be individualized based upon clinical and laboratory conditions. These, together with the common occurrence of CMV and EBV infection in high-risk patients, should direct the search for optimal preventive strategies for CMV and EBV infection occurrence after liver transplants.
Materials and Methods
A total of 344 consecutive pediatric patients (< 18 y) that underwent LDLT at Kyoto University Hospital between 1998 and 2011 were included in our study. This protocol conforms to the ethical guidelines of the 1975 Helsinki Declaration. Due to the retrospective nature of this study, Institutional Review Board (The Ethical Committee of the Faculty of Medicine at Kyoto University) approval was not needed.
Surgical procedure and immunosuppression therapy
The surgical techniques and preoperative management of transplant recipients at our center have been previously described in detail.9 In brief, the basic immunosuppression regimen consists of tacrolimus and low-dose corticosteroids.10 Oral tacrolimus was given every 12 hours starting 1 day before the operation, at dosages dependent on target trough levels of 10 to 15 ng/mL for the first 2 weeks and 5 to 10 ng/mL for the next 2 months. A dosage of 10 mg/kg of methylprednisolone was administered after reperfusion of the grafts and then 1 mg/kg twice a day for the first 3 days, 0.5 mg/kg twice a day for the next 3 days, and 0.3 mg/kg on day 7. From day 8, oral prednisolone (0.3 mg/kg/d) was added. In case of ABO-incompatible LDLT, additional immunosuppressive therapy was administered to inhibit humoral rejection as reported previously.11 Cyclophosphamide was orally administered at a dose of 2 mg/kg/d 7 days before surgery and was switched to 1 mg/kg/d of azathioprine from 1 month after surgery. In addition, prostaglandin E1 was given intravenously at a dosage of 0.01 mg/kg/min for 1 to 2 weeks after surgery. Corticosteroid pulse therapy was added weekly during the first month, and then gradually tapered during months 3 to 12 after LDLT. In case of rejection, a regimen of tacrolimus and steroid pulse therapy previously described was instituted.12,13
Postoperative antimicrobial prophylaxis
Antiviral prophylaxis, including ganciclovir, was not administered except in cases of seronegative recipients who received allografts from CMV-seropositive donors. Oral flomoxef, an oxacefem antibiotics were started from the first postoperative day and continued for 72 hours thereafter to prevent bacterial infection. Trimethoprim and sulfamethoxazole were administered once daily as a prophylaxis against pneumocystis. Oral miconazole was administered for 7 days after transplant for antifungal prophylaxis.
Diagnosis of viral infections
Cytomegalovirus and EBV status were evaluated weekly by serology and polymerase chain reaction testing. Viral infections were diagnosed based upon clinical findings, positive results from serologic tests, and liver biopsies.14 Infection was confirmed by detection of the viral DNA fragment in urine, buffy coat, ascites, and tissue samples using polymerase chain reaction.15 Any asymptomatic viral infection determined by serologic testing was excluded from the study. In addition, systemic lymphadenopathy was assessed, and biopsies of the lymph nodes at the body surface, including the axillary lymph nodes and inguinal lymph nodes, were performed to rule out PTLD.
Risk factors for infections
We evaluated the risk potential of several preoperative, operative, and postoperative variables. The preoperative variables included age at the time of LDLT, gender, clinical status, and the presence of preoperative infection and ascites. A standardized height score (SD score) was calculated for each patient as follows: Z score = (measured value – average value in normal population)/SD of the normal population. Laboratory variables included white blood cell count, C-reactive protein, electrolytes (calcium, phosphorus, magnesium, potassium and sodium), liver function markers (aspartate transaminase, alanine transaminase, total bilirubin, and albumin), coagulation factors (international normalized ratio, antithrombin ІІІ, platelet number, and prothrombin time), renal function markers (blood urea nitrogen and creatinine), preoperative hospital stay, ABO mismatching, and pediatric end-stage liver disease score. Operative variables included operation time, cold ischemic time, warm ischemic time, blood or blood product transfusion (packed red blood cells, fresh frozen plasma, and platelets), fluid transfusion (5% albumin), blood loss, and graft recipient body weight ratio. Postoperative variables included surgical complications (intra-abdominal hemorrhage, bile leak, and intestinal perforation), repeat surgery, postoperative intensive care unit stay, renal dialyses, length of insertion of intravascular catheter, intra-abdominal drainage, bile drainage, and graft dysfunction. Graft dysfunction was defined as persistent abnormal liver function with serum aminotransferase levels 2 to 3 times normal, with or without elevated bilirubin, and abnormal biopsy finding. Postoperative days were divided into 2 equal groups. Data of concomitant rejection or administration of steroid pulse therapy at the time of infection were also evaluated.
Statistical analyses were performed using the Statistical Package for the Social Science (SPSS: An IBM Company, version 16.0, IBM Corporation, Armonk, NY, USA). Data are presented as mean ± standard deviations, median, range, or percentage where appropriate. Cox proportional hazard regression model was used to evaluate the effect of different factors on the risk of CMV or EBV infections. Factors with a P value ≤ .05 were included in the multivariate analysis. A t test for quantitative data and chi square test for qualitative data were used to compare differences between LDLT recipients who did and did not develop either CMV or EBV viral infections.
Among 344 subjects who underwent LDLT from November 1998 to April 2011, forty-four required a second transplant, and 1 required a third. Of the study participants, 193 were female (56.1%). The mean age at the time of transplant was 3.95 ± 4.75 years (median, 1.38 y; range, 0.07-17.87 y). Indications for LDLT among patients were biliary atresia (n = 215, 62.6%), acute liver failure (n = 38, 11%), metabolic liver disease (n = 21, 6.1%), tumors (n = 17, 4.9%), Alagille syndrome (n = 13, 3.8%), liver cirrhosis/hepatitis (n = 8, 2.3%), Byler disease (n = 4, 1.2%), and others (n = 28, 8.1%). The mean follow-up duration was 7.1 ± 3.6 years (range, 0.02-13.20 y). Among all recipients and patients the preoperative serology identified 277 CMV cases and 272 EBV cases. A total of 156 patients (45.2%) developed viral infections, of those, 91 (26.5%) developed CMV, at 39.3 ± 34.6 days (median, 33 d; range, 3-300 d). Hepatitis was confirmed by liver biopsy in 13 cases, 9 developed fevers of unknown cause, 2 developed pneumonia, 1 was diagnosed as having CMV enteritis, and 1 patient manifested with pancytopenia. The remaining cases were diagnosed using polymerase chain reaction testing. A serologic combination of seropositive donor (D+) and seronegative recipient (R-) for CMV was reported in 69 patients, 14 of which developed CMV. Moreover, 55 of 85 patients that developed CMV were seropositive for CMV preoperatively. The incidence of CMV infection was significantly influence by the preoperative CMV seropositive status (P = .008) and a significant trend was identified for the serologic D+/R- combination (P = .094).
A total of 93 patients (27%) developed EBV at 3.99 ± 3.67 years (median, 2.93 y; range, 0.02-10.92 y) including 4 who developed PTLD, 9 that developed hepatitis, 8 that developed fever, and 2 that developed pneumonia. A serologic combination of seropositive donor (D+) and seronegative recipient (R-) for EBV was reported in 82 patients who developed EBV, and 39 patients were seropositive for EBV preoperatively. The incidence of EBV infection was not significantly influence by the preoperative EBV serologic state (P = .356) or D+/R- match (P = .432).
A total of 29 cases having both CMV and EBV infection were not included in risk factor analysis. The exact pattern of CMV and EBV infected patients is shown in Table 1.
Risk factors for cytomegalovirus and Epstein-Barr virus
Univariate analysis for factors affecting CMV infection after living-donor liver transplant
We compared factors affecting CMV infections after LDLT with those affecting EBV infections. Correlation testing revealed that seropositivity (R+) for CMV (P = .017), preoperative low platelet count (P = .01), preoperative diagnosis of fulminant hepatic failure (FHF) (P = .024), and frequent attacks of rejection (P < .001) correlated with postoperative CMV infections (Table 2A). Young age at the time of transplant (< 1 y) (P = .001), prolonged preoperative hospital stay (> 7 d) (P = .033), preoperative ascites (P = .018), prolonged intravascular catheter insertion (P = .014), preoperative CMV (R+) (P < .001), CMV (D+/R-) (P = .024), low preoperative blood albumin level (P = .041), high preoperative total blood bilirubin level (P = .01), high preoperative white blood cell count (P < .001), short cold ischemic time (P = .002), larger graft recipient body weight ratio (P < .001), more fresh frozen plasma (P = .008), new cases compared to old cases (P < .001), and frequent attacks of rejection (P < .001) correlated with postoperative EBV infections (Table 2B).
Multivariate analysis for factors affecting infection after living-donor liver
Potential predictors further examined with multivariate analysis revealed that frequent attacks of rejection (hazard ratio [HR], 1.58; 95% CI: 1.14-2.18; P = .006) independently predicted postoperative CMV infections, while preoperative low platelet count showed a trend (HR, 1.0; 95% CI: 0.996-1.00; P = .088) (Table 3A).
Regarding EBV, CMV (R+) (HR, 1.76; 95% CI: 1.03-2.18; P = .038) short cold ischemic time (HR, 1.0; 95% CI: 0.99-1.00; P = .02), larger graft recipient body weight ratio (HR, 1.3; 95% CI: 1.00-1.73, P =.047), and new cases compared to old cases (HR, 2.27; 95% CI: 1.14-4.52, P = .019) independently predicted postoperative EBV infections (Table 3B).
In our cohort there were 47 deaths, 2 were due (4.3%) to EBV-related PTLD, 1 was due (2.2%) to EBV hepatitis, and 1 was due (2.2%) to CMV infection. A total of 50 patients had graft dysfunction, 16 of which had CMV and 12 had EBV infections. There were no significant differences among infected and noninfected cases (P = .213 and P = .369).
Different effect of pathology on cytomegalovirus/ Epstein-Barr virus infection
In our cohort, 37 cases were diagnosed as having FHF. The incidence of CMV/EBV infection in these patients was compared to those in patients with chronic liver failure. The results showed that CMV was detected in 40.5% of FHF patients and 24.8% of chronic liver failure patients (P = .035), while EBV was detected in 35.1% of FHF patients and 26.1% of chronic liver failure patients (P = .20).
The contrast pattern of cytomegalovirus/Epstein-Barr infection in pediatric
living-donor liver transplant over the 13-year period
Studying the patterns of CMV and EBV infections from 1998 to 2011, we found that the frequency of EBV but not CMV was lower in the older cases and gradually increased since 2002.
The results for pediatric liver transplants improved dramatically after development of effective immunosuppressive drugs in the second half of the 1980s.16 Despite the antiviral strategies of today, CMV and EBV infection have remained the most troublesome complications after LDLT.6,17,18 Our results showed that 155 of pediatric patients (45.2%) had at least 1 viral infection episode after LDLT. A diagnosis of CMV infection was confirmed in 91 of these patients (26.5%), which is consistent with a previous report from our center.8 Other centers showed a higher incidence of infection, reporting that CMV affects up to 75% of liver transplant recipients directly or indirectly.17-19 A possible explanation for the discrepancy between our results and theirs is the lower target immunosuppressive trough level at our center compared with others, which did not lead to impairment of the immune system.
On the other hand, EBV infection was diagnosed in 93 of our patients (27%), which is considered high compared to Shepherd and his colleagues.19 We attribute this increase to the difference in our inclusion criteria, as we included all infected cases regardless of severity, while Shepherd and accociates19 included only severe cases. Moreover, our follow-up was longer than theirs was. Earlier reports by Singh and accociates20 described data close to ours. Altogether, our findings may suggest that although the incidence of viral infection at our center was relatively high compared to some centers,19,21,22 we had a lesser frequency of CMV disease and PTLD compared to a previous report.19 These results may be explained by the serial real-time polymerase chain reaction screening protocol used at our center for evaluating the CMV and EBV genome that was found to be helpful for early detection and monitoring of viral infections for preemptive treatment of PTLD.23-25
Cytomegalovirus infection was diagnosed earlier than EBV infection at our center. A possible explanation for this is that CMV infection occurs due to a severely suppressed immune status in the recipient. Uncontrolled CMV proliferation in the absence of CMV-specific immune-protective CD8+ cells after transplant favors the occurrence of CMV infection.26
Moreover, the incidence of CMV infection was significantly influence by preoperative seropositivity for CMV, while EBV showed no significant difference based on preoperative seropositivity. This difference may be attributed to the fact that CMV infection occurs in the early period while the immune system is severely suppressed that allow CMV proliferation, while EBV usually occur later after reduction of immunosuppressive drugs.
The chance of a satisfactory outcome in the management of CMV and EBV infection depends on early diagnosis and the use of strategies aimed at prevention. Early treatment as well as the prevention of CMV and EBV infections in children may be facilitated by knowledge of the primary risk factors for these infections in this population. In the current study, numerous risk factors were identified as being associated with CMV. A recipient seropositive status was associated with CMV infection and Lejungman and accociates also reported that the seropositivity of patients is a major risk for CMV infection after transplant.27 We and others25,28 also found that FHF was a major risk factor for CMV infection, and although the mechanism of this association is not completely understood, FHF is associated with very high levels of tumor necrosis factor alpha which may directly promote viral replication.
Concomitant rejection was an independent risk factor for CMV infection and similar data have been reported previously.19,29 These data raise the possibility that immunosuppression is a modifiable risk factor for serious viral infection. Although it is difficult to assess which comes first, rejection or infection, avoiding over-immunosuppression by carefully monitoring immunosuppressant drug dosages and blood levels and limiting steroid use in pediatric patients should be an important goal.19 Fortunately, immunosuppression reduction is highly encouraged at our center. No less than 15% of pediatric patients (which is high compared with other centers) achieve complete withdrawal of immunosuppression and develop a state of operational tolerance.30
A low preoperative platelet count was found to be an independent risk factor for CMV. This can be explained by the clinical association of FHF in those patients. Another explanation is that a low platelet count itself may be an indication for more transfusions that may lead to viral infection transmission.
Regarding factors affecting EBV infections after LDLT, correlation testing showed a young age at the time of transplant (< 1 y) to be an associated risk for EBV infection. Guthery and his associates and others also reported a young age at transplant to be associated with an increased risk of EBV-PTLD in children undergoing liver transplants.19,31,32 A prolonged postoperative hospital stay also was found to be an associated risk factor for EBV, which can be explained by excessive exposure to hospital-acquired pathogenic microorganisms.33 We, as well as others, found preoperative CMV (R+) and CMV (D+/R-) to be risk factors for EBV infection.34-36 However, CMV (R+) but not CMV (D+/R-) was determined to be an independent risk factor for EBV infection that can be explained by the immunologic cascade triggered by the CMV viral infection, where for example release of tumor necrosis factor alpha plays a pathogenic role.37 These indirect effects of CMV may exert a state of enhanced immune suppression with a concurrent higher risk of EBV infection and hence PTLD. Therefore, effective prevention of CMV also may prevent EBV infections, primarily by limiting the effect of CMV on immune regulation.38 Hypoalbuminemia is a common problem among persons with acute and chronic liver disease at the time of hospital admission and serum albumin level is an important prognostic indicator. Among hospitalized patients, lower serum albumin levels correlate with an increased risk of morbidity and mortality. Our results showed that low preoperative serum albumin levels, more fresh frozen plasma, prolonged intravascular catheter insertion, preoperative ascites, and high preoperative serum total bilirubin level are associated risk factors for EBV, which may reflect that cases with preoperative poor medical conditions are more prone to infection with EBV after a liver transplant.
We also found elevated white blood cell counts to be associated with EBV infection and an abnormally elevated preoperative white blood cell count has been associated with postoperative morbidity and mortality.39 Frequent attacks of rejection was an independent risk for EBV, and similar data has been reported by Shepherd and accociates.19 This finding can indicate that EBV can be induced by cytokine-mediated interactions due to rejection with normal cells of the immune system because we could not find any relation between EBV infection and immunosuppression. Graft-recipient-body-weight ratio also was found to be an independent factor for EBV infection in our study. This can be explained by the fact that graft recipient body weight ratio differences are usually found in younger patients who were more prone to infection.31 Although, we found that new cases compared to old were independently predictive of EBV infection, we found that graft survival is better in recent cases compared with old cases that may be due to the frequent monitoring and early detection of EBV infection.
Our findings suggest that although we have a relatively low incidence of infection at our center, viral infections still remain the second cause of death after sepsis in pediatric patients following LDLT. Similar findings on the effect of CMV on mortality rates have been reported in liver transplant recipients.4 In our study, EBV related PTLD lead to the death of 2 patients. Mortality due to PTLD after pediatric solid-organ transplant was previously reported to be 12% to 32%40,41 and consist with our findings this incidence has been significantly reduced by EBV DNA monitoring and tapered immunosuppressive drug regimens.
Altogether, we noticed that the number of patients in our study that were infected and those that died as a result of CMV and EBV was lower than that in other centers which perform LDLT. This discrepancy can be due to many factors including the efficacy of our EBV screening program in decreasing the incidence of EBV-related PTLD as well as the fact that prophylactic antiviral therapy in D+/R- combination cases may result in better outcomes at our center. In addition, the lower target level of immunosuppression at our center and the weaning protocol in which physicians start the process of gradual decreasing immunosuppressive drugs until finally stopping them completely appears to be beneficial.
In conclusion, based on the results of the current study, it is strongly recommended that early diagnosis of CMV and EBV is essential for LDLT patients. In this regard, CMV and EBV quantitative polymerase chain reaction testing may aid in surveillance. Such surveillance should be extended, especially in the case of infants who have a bigger graft recipient body weight ratio, those with a previous history of FHF, and patients at a particularly high risk of CMV and EBV infection. Additional research should be directed at improving the sensitivity, specificity, and predictive values of polymerase chain reaction and other testing modalities for this purpose. Multicenter trials are encouraged to evaluate the role of various regimens in the prevention of CMV and EBV infection in high risk patients.
Volume : 13
Issue : 1
Pages : 75 - 82
DOI : 10.6002/ect.mesot2014.O26
From the 1Department of Microbiology and Immunology, Graduate
School of Medicine, Assiut University, Assiut, Egypt; the 2Department
of Hematology and Immunology, Graduate School of Medicine, Umm Al-Qura
University, Makkah, Saudi Arabia; the 3Department of Anesthesia,
Graduate School of Medicine, Assiut University, Assiut, Egypt; the 4Department
of Anesthesiology, King Abdullah Medical City, Makkah, Saudi Arabia; and the
5Department of Hepatobiliary Pancreatic Surgery and Transplantation,
Graduate School of Medicine, Kyoto University, Kyoto, Japan
Acknowledgements: Hanaa Nafady-Hego, Hamed Elgendy, and Shinji Uemoto carried out the research. Hanaa Nafady-Hego, Hamed Elgendy, and Shinji Uemoto wrote the paper. Hanaa Nafady-Hego, Hamed Elgendy, and Shinji Uemoto participated in research design. Hanaa Nafady-Hego and Hamed Elgendy conducted the data analysis. Shinji Uemoto directed the transplant program. All authors have no potential interest to declare.
Corresponding author: Hanaa Nafady-Hego, MD, PhD, Hepatobiliary Pancreatic Surgery and Transplantation, Graduate School of Medicine, Kyoto University, 54 Kawara-cho, Shogoin, Sakyo-ku, 606-8507, Kyoto, Japan
Phone: +81 75 751 4328
Fax: +81 75 751 4328
E-mail: email@example.com , firstname.lastname@example.org
Table 1. Characteristics of Cytomegalovirus, Epstein-Barr Virus, and Noninfected Cases After Pediatric Living Donor Liver Transplant
Table 2A. Univariate Analysis of Risk Factors for Cytomegalovirus Infection in 344 Pediatric Living Donor Liver Transplant Recipients
Table 2B. Univariate Analysis of Risk Factors for Epstein-Barr Virus Infection in 344 Pediatric Living Donor Liver Transplant Recipients
Table 3A.Multivariate Analysis of Factors Affecting CMV Infection in 344 Pediatric LDLT Recipients
Table 3B. Multivariate Analysis of Factors Affecting EBV Infection in 344 Pediatric LDLT Recipients