Mannose-binding lectin (MBL) is a serum C-type lectin belonging to the collectin protein family. Its main function is to bind carbohydrates present on the surface of a wide variety of pathogens, including bacteria, fungi, viruses, and parasites.1 As a component of the innate immune system, MBL generates a rapid response against pathogens during the lag phase needed for adaptive immunity to become operative.2 In serum, MBL is present in an oligomeric form (mainly tetramers and hexamers) bound to MBL-associated serine proteases (MASPs), among which MASP-2 plays the most important role in immune response.3 Upon binding to pathogens, the MBL/MASP-2 complexes mediate direct opsonophagocytosis4 as well as complement activation via the formation of the classical C3 convertase through the so-called lectin pathway.5
Certain polymorphisms in MBL2, the gene responsible for encoding MBL, are associated with low levels of MBL in serum.6 Three single nucleotide polymorphisms at codons 52, 54, and 57 of exon 1 (named the D, B, and C variants, respectively) are major determinants of serum MBL levels and are collectively called variant haplotypes (O), whereas the wild-type haplotype is referred to as A.7, 8 Additional single nucleotide polymorphisms at positions −551 (H/L), −221 (X/Y), and +4 (P/Q) in the 5′-flanking region also influence serum MBL levels.9 Homozygous carriers for wild-type MBL2 have high MBL levels in serum, and those homozygous for exon 1 variant genotypes have very low MBL serum levels.
Since the first descriptions of a clinical association of low serum levels of MBL with meningococcal disease,10 many other studies have found an association between low levels of MBL or variant MBL2 genotypes and a wide variety of infectious diseases.1
MBL is mainly produced in the liver, but low amounts of the protein have been found in the small intestine and testis tissue.11 After liver transplantation, the MBL serum levels depend on the donor MBL2 genotype. Bouwman et al.12 found that serum MBL levels decreased dramatically in recipients with a pretransplant wild-type MBL2 genotype after they received liver grafts from donors with a variant MBL2 genotype. In this study, recipients of liver grafts from donors with a variant MBL2 genotype had a higher risk of infection. However, the number of patients studied was low, and the findings need further confirmation. A very recent study by Worthley et al.13 demonstrated a 3-fold greater likelihood of a clinically significant infection developing in recipients of a liver transplant from donors with exon 1 coding mutations. However, the severity and outcome of infections according to the MBL2 genotypes were not evaluated in that study.
The aim of our study was to evaluate the influence of donor and recipient MBL2 genotypes on the incidence and outcomes of infections after liver transplantation.
aHR, adjusted hazard ratio; CI, confidence interval; CMV, cytomegalovirus; HR, hazard ratio; ICU, intensive care unit; IL, interleukin; IQR, interquartile range; MASP, mannose-binding lectin–associated serine protease; MBL, mannose-binding lectin; MELD, Model for End-Stage Liver Disease; OR, odds ratio; PCR, polymerase chain reaction; SD, standard deviation; UTR, untranslated region.
PATIENTS AND METHODS
From January 2004 to December 2006, all patients undergoing liver transplantation were evaluated at the Hospital Clinic of Barcelona, an 850-bed tertiary care hospital in Spain. The availability of a whole-blood sample from the donor was required for inclusion.
All patients received intravenous steroids in the peritransplant period and calcineurin inhibitors (cyclosporin or tacrolimus). Steroids were tapered progressively until discontinuation at approximately 6 months post-transplantation.
Prophylactic Schemes for Infections
All patients received a presurgical dose of 2 g of ceftazidim plus 400 mg of teicoplanin. Prophylaxis for Pneumocystis jiroveci with trimethoprim-sulfamethoxazole (160 mg/800 mg) was administered daily during the first year post-transplantation. Antifungal prophylaxis with fluconazole was administered to patients requiring prolonged antibiotic therapy (more than 3 days). In patients at high risk for aspergillosis (acute renal failure, requirement for renal replacement therapy, reoperation in the early posttransplantation period, or a positive galactomannan antigen test), intravenous liposomal amphotericin B (3 mg/kg/day) was given for a period of not less than 1 week. In patients with a donor-recipient cytomegalovirus (CMV) mismatch (D+/R−), oral valganciclovir (450 mg twice daily, adjusted by renal function) was given for 100 days. In the case of positive pp65 antigenemia in the follow-up (more than 30 positive cells), preemptive treatment with oral valganciclovir adjusted to renal function (at an equivalent dose of 900 mg twice daily with normal renal function) was administered for 14 days. All patients were monitored for CMV pp65 antigenemia weekly for the first 2 months and twice monthly 6 months after transplantation.
Infectious Event Definitions
Guidelines from the Centers for Disease Control and Prevention were followed for nosocomial bacterial infections.14 All bacterial infection definitions have been published elsewhere.15 Briefly, the presence of an infectious episode was established with the following criteria:
1A positive culture of a pathogenic microorganism (Mycobacterium tuberculosis or Legionella pneumophilia) in any sample.
2Isolation of any microorganism from a sample obtained under sterile conditions.
3Isolation of a potentially pathogenic microorganism in a sample from any location accompanied by compatible symptoms of infection.
4An infection (suggested by clinical data) without microbiological isolation that was resolved completely with antimicrobial treatment.
Febrile episodes were excluded from the analysis if no microorganism was isolated and no antimicrobial treatment was given with clinical resolution of the episode or if a definitive diagnosis was provided that precluded an infectious agent. We included only those infectious events considered to be major infections: bloodstream infections, pneumonia (upper respiratory tract infections were excluded), deep surgical wound infections, febrile urinary infections (a diagnosis of uncomplicated cystitis was excluded), and other infections requiring hospital admission.
CMV disease was defined according to the guidelines proposed by Ljungman et al.16 CMV viral syndrome was diagnosed when compatible clinical symptoms were present and evidence of CMV replication in the blood was detected, either by the presence of a positive pp65 antigenemia test or by positive DNAemia. CMV end-organ disease was diagnosed only when a biopsied specimen revealed typical histopathological findings of CMV infection and positive immunohistochemical staining. Fungal infections were categorized as definite or probable on the basis of the European Organization for Research and Treatment of Cancer/Mycoses Study Group definitions.17 Septic shock was diagnosed in patients with systemic inflammatory response syndrome and persistent dysfunction of at least 1 organ caused by hypoperfusion requiring hemodynamic support (fluids and vasoactive drugs). Systemic inflammatory response syndrome was diagnosed by the presence of 2 of the following: temperature > 38°C or < 36°C, heart rate > 90 ppm, tachypnea (>20 rpm, hyperventilation, or PCO2 < 32 mm Hg), and altered white cells (>12,000 or <4000 leukocytes per cubic millimeter or >10 nonsegmented neutrophils on a differential count).18, 19
After hospital discharge, patients were visited at least once monthly. On the appearance of fever or symptoms of infection, patients were hospitalized. On readmission, a detailed physical examination, standard laboratory tests, a chest X-ray, and blood and urine cultures (or sterile fluids when appropriate) were performed.
Genomic DNA was extracted from 1.5-mL ethylene diamine tetraacetic acid–treated whole blood samples with the QIAamp DNA blood minikit according to the manufacturer's instructions (QIAGEN GmbH, Hilden, Germany) and stored at −20°C until it was used.
Genotyping of MBL2 was done by a polymerase chain reaction (PCR)/sequence-based typing technique, as previously reported.20, 21 Briefly, a 969-bp fragment of MBL2 encompassing a region from the promoter to the end of exon 1 was obtained by PCR amplification with the sense 5′-GGGGAATTCCTGCCAGAAAGT-3′ and antisense 5′-CATATCCCCAGGCAGTTTCCTC-3′ primers and the Expand 20kbPLUS PCR system (Roche Diagnostics GmbH, Mannheim, Germany). The resulting PCR product was treated with ExoSAP-IT (USB Corp., Cleveland, Ohio) and then subjected to direct cyclic sequencing with the BigDye Terminator version 1.1 cycle sequencing kit (Applied Biosystems, Warrington, United Kingdom) according to the manufacturer's instructions with the aforementioned sense and antisense gene-specific primers. Sequencing reactions were analyzed on an automated capillary DNA sequencer (ABI-Prism 3100 genetic analyzer, Applied Biosystems).
Genetic groups were defined according to the phenotypic correlates with MBL serum levels.6, 22 Donors were categorized to have sufficient (A/A) or insufficient (A/O, O/O) MBL levels according to the structural alleles defined by polymorphic positions at exon 1 of the MBL2 gene [A designates the wild-type allele, and O designates any of the 3 structural mutations (B, C, and D) reported at exon 1]. For promoter and untranslated region (UTR) polymorphisms, the following groups were defined: (1) allele H/L (promoter −550) high producers (HH) and low producers (HL or LL), (2) allele P/Q (5′-UTR+4) high producers (QQ) and low producers (PP and PQ), and (3) allele Y/X (promoter −221) high producers (YY and YX) and low producers (XX).22, 23 On the basis of the study by Worthley et al.,13 the combinations of the promoter −220 and exon 1 haplotypes were evaluated, with YA/YA and YA/XA defined as sufficient MBL production and the rest of the combinations defined as insufficient MBL production.
For sample size estimates, we assumed that 50% of liver transplant recipients had bacterial infections. Under the assumption that MBL2 variant genotypes (50% of the patients) increase the risk of bacterial infection up to 70% in recipients, we required 94 patients to detect the aforementioned difference with a power of 80% at a significance level of 95%.
For statistical analysis, we used the SPSS (version 12.0, SPSS, Inc., Chicago, IL) and Stata 9.1 (StataCorp LP, College Station, TX) statistical packages. Continuous variables were summarized as means [standard deviation (SD)] or as medians (interquartile range or range) according to their homogeneity. Categorical variables were compared with the χ2 test or Fisher's exact test when appropriate. Continuous variables were compared with the Mann-Whitney U test or Student T test according to their homogeneity. For regression models, we included predictors with P < 0.30 in the univariate analysis. An analysis of factors independently associated with bacterial infection was performed with a logistic regression model. An analysis of factors independently associated with mortality or graft loss was performed with the Cox regression model. Associations are given as odds ratios or hazard ratios with a confidence interval (CI) established at 95%. A 2-sided probability value < 0.05 was considered to be significant.
During the 3-year period, 244 liver transplants were performed in our center, and a whole-blood sample was available for 107 donors. In 12 cases, we were unable to amplify DNA; therefore, 95 liver transplant patients were finally analyzed.
The mean age of the 95 recipients was 53.9 years (SD, 10.9 years), 59 being men (62%). The most frequent reason for transplantation was liver cirrhosis due to hepatitis C virus (54.5% or 45 cases, 19 of which had hepatocellular carcinoma); this was followed by alcoholic cirrhosis (14 cases) and fulminant hepatic failure (11 cases). The mean Model for End-Stage Liver Disease score was 15.6 (SD, 7.4). The incidence of bacterial infections was 36% (34 patients), the incidence of fungal infection was 12% (11 patients), and the incidence of CMV disease was 11% (10 patients). Noninfectious posttransplant complications included 27 cases of acute rejection, 7 cases of dialysis requirement, and 22 cases of major reoperation.
Thirteen patients died during follow-up, and 1 required retransplantation for severe rejection. Infection-related death was the most frequent diagnosis and was the cause of 11 of the 13 deaths (85%). Infectious causes of death were septic shock (6 cases), invasive aspergillosis (3 cases, with 1 associated with cerebral hemorrhage), nosocomial pneumonia (1 case), and surgical wound infection (1 case). Two patients died because of noninfectious events: multi-organ dysfunction syndrome in one case and cerebral edema in the other.
The frequencies of donor MBL2 haplotypes are summarized in Table 1.
Table 1. Frequencies of Donor Mannose-Binding Lectin 2 Genotypes
Abbreviation: UTR, untranslated region.
Exon 1 (−52, −54, and −57)
The pathogens responsible for posttransplant infections are summarized in Table 2. Bacterial infections were the most frequent in the cohort, especially those caused by enteric gram-negative bacilli and nonfermentative gram-negative bacilli. Only 10 patients developed CMV disease. Of the 11 fungi isolated, Aspergillus spp. were the most frequent pathogens.
Table 2. Pathogens Responsible for Infections in the Cohort of Liver Transplant Recipients
Number of Isolations
NOTE: In 10 patients with the diagnosis of bacterial infection, no microbiological diagnosis was achieved (5 had pneumonia, and 5 had acute cholangitis).
We found no differences in the incidence of bacterial and fungal infections according to the donor MBL2 genotypes (Table 3). The incidence of CMV disease was similar among genetic groups (no statistically significant differences). We also analyzed the incidence of fungal infection according to coding mutations, and no statistical differences were found.
Table 3. Incidence of Infection and Graft/Patient Survival According to the Donor MBL2 Genotype
Patients receiving a graft from a donor with exon 1 MBL–low-producing genotypes had a lower survival rate (Table 3 and Fig. 1). This lower survival rate in recipients of a liver with a coding mutation in exon 1 was due to infections of higher severity (deaths were mainly due to septic shock). For this reason, we performed binary logistic regression to find independent variables associated with the development of posttransplant bacterial infections (Table 4). Only posttransplant reoperation and requirement of renal replacement therapy were independent predictors of bacterial infections after transplantation.
Table 4. Risk Factors for Bacterial Infection After Liver Transplantation
Multivariate Analysis: Adjusted OR (95% CI)
Bacterial Infection During Follow-Up (n = 34)
Absence of Bacterial Infection During Follow-Up (n = 61)
NOTE: Although the donor MBL2 variant genotype had a P value > 0.30, it was included in the model for theoretical reasons.
Abbreviations: CI, confidence interval; CMV, cytomegalovirus; IQR, interquartile range; MBL2, mannose-binding lectin 2; MELD, Model for End-Stage Liver Disease; OR, odds ratio; SD, standard deviation.
Donor exon 1 MBL2 coding mutation
Mean recipient age (SD)
Mean pretransplant MELD (SD)
Median cold ischemia time (IQR)
Median surgical time (IQR)
CMV serology: D+/R−
Creatinine > 1.7 mg/dL at first month
Requirement for dialysis
When clinical and biological characteristics of bacterial infection episodes were compared according to donor exon 1 MBL2 genotypes (Table 5), septic shock was statistically more frequent in patients receiving an exon 1 MBL2 variant genotype liver graft (46% versus 11%, P = 0.004). The serum levels of C-reactive protein and creatinine were significantly higher in recipients of a liver graft with an exon 1 MBL2 variant genotype during a bacterial infection episode (P = 0.019 and P = 0.039, respectively). For microbiologically confirmed episodes, there was no difference in gram-positive bacterial infections versus gram-negative bacterial infections according to the MBL2 exon 1 donor genotype (P = 0.160).
Table 5. Differences in Clinical and Analytical Parameters in Bacterial Infection Episodes According to the Donor MBL2 Exon 1 Genotypes
Donor MBL2 Exon 1 Sufficient (A/A; Number of Episodes = 45)
Donor MBL2 Exon 1 Insufficient (A/O, O/O; Number of Episodes = 22)
Abbreviations: ICU, intensive care unit; MBL2, mannose-binding lectin 2; SD, standard deviation.
A Cox regression analysis for liver transplant failure (mortality or retransplantation) showed that a donor exon 1 MBL2 variant genotype [adjusted hazard ratio (aHR), 9.64; 95% CI, 2.59-36.0], the Model for End-Stage Liver Disease score (aHR, 1.14; 95% CI, 1.05-1.23), and bacterial infections (aHR, 11.1; 95% CI, 2.73-44.9) were independent variables associated with worse outcomes (Table 6).
Table 6. Cox Proportional Hazards Model Regression Analysis of Variables Associated with Worse Outcomes of Liver Transplantation
Crude HR (95% CI)
Adjusted HR (95% CI)
Abbreviations: CI, confidence interval; CMV, cytomegalovirus; HR, hazard ratio; ICU, intensive care unit; MBL2, mannose-binding lectin 2; MELD, Model for End-Stage Liver Disease.
Donor MBL2 exon 1 insufficient (A/O, O/O)
Length of ICU admission
The 2 main findings in our study are (1) MBL2 genotypes are not predictive for posttransplant infections and (2) infections are more severe in recipients of a liver transplant from a donor with an exon 1 MBL2 variant genotype, with infections being an independent variable associated with worse outcomes.
On the basis of the findings of the present study and a previous study by our group in renal transplant recipients, MBL2 genotypes do not seem to be predictive for bacterial infections in transplant recipients in our population.15 MBL is able to recognize a wide range of pathogens; however, the grade of binding to the surface of bacteria varies from high binding for Staphylococcus aureus or beta-hemolytic streptococci to low binding for Streptococcus pneumoniae.24 In the case of enteric bacilli or Pseudomonas aeruginosa, the binding is heterogeneous. In our cohort of patients, the most frequently isolated bacteria were those with low-grade binding to MBL. Thus, the influence of low MBL serum levels on the incidence of certain infections may depend on the flora responsible for posttransplant infections. Another relevant issue is that most research on MBL deficiency and associated infectious diseases is focused on a sole pathogen, such as Neisseria meningitidis25–27 or S. pneumoniae.28 In the case of infectious syndromes caused by a wide range of pathogens, these associations may be difficult to demonstrate. This could be the reason for the differences between the 2 previous studies in liver transplant recipients12, 13 and ours regarding the impact of MBL2 coding mutations on the incidence of posttransplant infections.
Two previous investigations, including one by our group, have shown an association of MBL levels29 and MBL2 genotypes15 with CMV disease after organ transplantation. However, a recent article by Verschuren et al.30 did not find an association between MBL serum levels and CMV disease in recipients of simultaneous kidney-pancreas transplantation. In our cohort of liver transplant recipients, no association with CMV disease was found according to the donor genotype. The low incidence of CMV disease (10 cases) in our study limits the strength of this conclusion, and more studies are needed in this field.
Regarding fungal infections, a previous study found that MBL2 variant genotypes were associated with invasive fungal infections in patients with allogeneic stem cell transplantation.31 In fact, Aspergillus spp. and Candida spp. induce strong binding to MBL.24 The lack of differences in the incidence of fungal infections according to MBL2 genotypes was probably due to the low number of cases included.
Donor exon 1 MBL2 variant genotypes were associated with worse outcomes in liver transplant recipients, mainly because of infections of greater severity. When we compared episodes of bacterial infection according to the MBL2 genotype, patients with a liver graft from a donor with an exon 1 MBL2 variant genotype developed septic shock in almost 50% of the episodes, whereas recipients of wild-type grafts developed shock in 11% of the episodes. This could be explained by the ability of MBL to modulate the production of proinflammatory cytokines. In one study using an ex vivo model, the addition of high concentrations of MBL to the blood of MBL-deficient donors profoundly decreased the production of interleukin-6 (IL-6), IL-1β, and tumor necrosis factor alpha by monocytes in response to meningococci, whereas lower concentrations of MBL increased IL-6 and IL-1β production.32 These proinflammatory cytokines are responsible in part for the initiation of the cytokine storm and, thereafter, the classical severe sepsis and septic shock syndromes.33 The association of variant genotypes and worse outcomes and development of severe sepsis and septic shock in adult and pediatric patients has been found in previous studies.34–39 Other investigations have suggested that heterozygosity in the MBL2 genotype could represent a protective factor for worse outcomes in critically ill patients.40 Although we did not determine the proinflammatory cytokines during the infectious events, we found that levels of C-reactive protein were higher in patients with a liver transplant from a donor with an exon 1 MBL2 variant genotype. Since the early 1990s, a positive correlation has been found between the serum levels of tumor necrosis factor alpha and IL-6 and C-reactive protein. Higher levels of these 3 biomarkers are associated with worse prognosis of sepsis.41
If these results are confirmed by larger studies, this may have several implications for liver transplantation. One would be to increase the length of the antibiotic prophylaxis in recipients of an exon 1 MBL2 variant donor. However, this may increase the incidence of pathogens resistant to multiple antibiotics.42 A future option could be the use of recombinant MBL substitutive treatment in recipients of a liver transplant from a donor with an exon 1 MBL2 variant genotype.43 As the critical period for life-threatening infections after liver transplantation is well known (around the first 3 months post-transplantation), substitutive therapy could be limited to this period. Finally, donor MBL2 genotypes should be screened in studies evaluating the severity of posttransplant infections as they could represent a major risk factor for infection-related mortality.
Our study has several limitations, and the results must be interpreted with caution. The sample size was estimated not to evaluate the outcome of liver transplantation but to estimate the differences in the incidence of infections. When single-gene variants are studied, environmental factors are usually responsible for the higher risk of the dependent variable. In our case, bacterial infections were the most important risk in the prognosis of liver transplantation. Thus, a confirmation of this association in a higher number of patients is needed to avoid type I errors.
In conclusion, the findings of the present study indicate that liver transplantation from an exon 1 MBL2 variant genotype donor is not associated with a higher incidence of infections but significantly influences the outcome, mainly because of infectious events of higher severity. Larger studies are needed to further confirm these findings.