Hypocomplementemia in Kidney Transplant Recipients: Impact on the Risk of Infectious Complications

Authors

  • M. Fernández-Ruiz,

    Corresponding author
    • Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • F. López-Medrano,

    1. Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • P. Varela-Peña,

    1. Department of Immunology, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • J. M. Morales,

    1. Department of Nephrology, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • A. García-Reyne,

    1. Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • R. San Juan,

    1. Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • C. Lumbreras,

    1. Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • D. Lora-Pablos,

    1. Unit of Clinical Research, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    2. CIBER de Epidemiología y Salud Pública (CIBERESP)
    Search for more papers by this author
  • N. Polanco,

    1. Department of Nephrology, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • A. Andrés,

    1. Department of Nephrology, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • E. Paz-Artal,

    1. Department of Immunology, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author
  • J. M. Aguado

    1. Unit of Infectious Diseases, Instituto de Investigación Hospital “12 de Octubre” (i+12), Hospital Universitario “12 de Octubre”, School of Medicine, Universidad Complutense, Madrid, Spain
    Search for more papers by this author

  • This study was partially presented at the 50th Annual Interscience Congress on Antimicrobial Agents and Chemotherapy (ICAAC), Boston (September 12–15, 2010) and at the 24th International Congress of The Transplantation Society (TTS), Berlin (July 15–19, 2012).

Corresponding author: Mario Fernández-Ruiz

mario_fdezruiz@yahoo.es

Abstract

The usefulness of monitoring of complement levels in predicting the occurrence of infection in kidney transplant (KT) recipients remains largely unknown. We prospectively assessed serum complement levels (C3 and C4) at baseline and at months 1 and 6 in 270 patients undergoing KT. Adjusted hazard ratios (aHRs) for infection in each posttransplant period were estimated by Cox regression. The prevalence of C3 hypocomplementemia progressively decreased from 21.5% at baseline to 11.6% at month 6 (p = 0.017), whereas the prevalence of C4 hypocomplementemia rose from 3.7% at baseline to 9.2% at month 1 (p = 0.004). Patients with C3 hypocomplementemia at month 1 had higher incidences of overall (p = 0.002), bacterial (p = 0.004) and fungal infection (p = 0.019) in the intermediate period (months 1–6). On multivariate analysis C3 hypocomplementemia at month 1 emerged as a risk factor for overall (aHR 1.911; p = 0.009) and bacterial infection (aHR 2.130; p = 0.014) during the intermediate period, whereas C3 hypocomplementemia at month 6 predicted the occurrence of bacterial infection (aHR 3.347; p = 0.039) in the late period (>6 month). A simple monitoring strategy of serum C3 levels predicts the risk of posttransplant infectious complications in KT recipients.

Abbreviations
APN

acute pyelonephritis

AR

acute graft rejection

CI

confidence interval

CMV

cytomegalovirus

CPR

C-reactive protein

ESRD

end-stage renal disease

HGG

hypogammaglobulinemia

HLA

human leukocyte antigen

HR

hazard ratio

HSV

herpes simplex virus

ICC

interclass correlation

Ig

immunoglobulin

INR

internationalized normal ratio

KT

kidney transplant

LRTI

lower respiratory tract infection

MBL

mannose-binding lectin

MDRD

modification of diet in renal disease

OR

odds ratio

SD

standard deviation

SOT

solid organ transplant

VVZ

varicella-zoster virus

Introduction

The complement system plays a crucial role in the innate immune response against infection and enhances the adaptive immunity by means of an intricate interaction between soluble activation products and cell surface receptors. Activation of the complement system triggers a constellation of effects including opsonization, chemotaxis, phagocytosis and cytotoxicity [1]. These host-defense functions become particularly relevant in the setting of immunosuppression after solid organ transplantation (SOT), which is mostly targeted against the adaptive immune arm [2]. The lectin pathway—one of the three complement activation cascades—has been extensively studied in the SOT population, thus demonstrating that both mannose-binding lectin (MBL) serum levels and polymorphisms of the MBL gene (mbl2) exert an important impact on the risk of infection in liver [3-5] and kidney transplant (KT) recipients [6-8]. Nevertheless, the assessment of MBL status is still far from being incorporated into routine clinical practice.

A more feasible approach to explore the complement system in SOT recipients could be based on the monitoring of a number of serum complement factors during the posttransplant period. The fourth component of the complement system (C4) holds a central position in the classical pathway, whereas all three activation cascades—classical, alternative and lectin pathways—converge at the level of the third component (C3) to form the C5 convertase (C4bC2aC3b from classical and MBL pathways, and [C3b]2Bb from alternative pathway), finally resulting in the formation of the membrane attack complex [1]. In addition, cleavage products of both factors potentiate antibody response, trigger local inflammatory responses, and enhance immunologic memory [9]. Although previous studies have found that decreased levels of C3 may be associated to the occurrence of infectious complications after liver [10] and heart transplantation [11], the evidence on this issue is scarce. This study was aimed at investigating the incidence and predisposing factors for hypocomplementemia in a large series of KT recipients, as well as to evaluate the usefulness of the monitoring of serum complement factors (C3 and C4) in predicting the risk of infection in this population.

Methods

Study population

We conducted a prospective cohort study at the University Hospital “12 de Octubre” (Madrid, Spain), a 1360-bed tertiary-care center with more than 25 years of experience in SOT. Beginning in November of 2008 all consecutive patients aged 18 years or older who underwent KT at our institution were included in a prospective immune status assessment based on serial measurements of total lymphocyte and peripheral blood lymphocyte subpopulations counts, serum immunoglobulin (Ig) levels (IgG, IgA and IgM), and serum complement levels. Partial results have been previously published [12]. We excluded patients with known pretransplant primary immunodeficiencies or human immunodeficiency virus infection, and those who died or developed graft loss within the first week. The local Clinical Research Ethics Committee approved the study protocol, and written informed consent was obtained from all participants.

Study design and immune status assessment

Serum samples were collected just before transplantation (baseline) and at months 1 and 6. Serum complement levels (C3 and C4) were determined by nephelometry (Immage 800, Beckman Coulter, Villepinte, France). Normal ranges, as established by our laboratory, were between 83.0 and 171.0 mg/dL for C3, and between 14.0 and 38.0 mg/dL for C4. Patients were enrolled at the time of transplantation and followed for at least one year, unless death or graft loss were documented earlier. We divided the posttransplant follow-up period in three intervals: early (first month), intermediate (months 1–6) and late (>6 months). The primary study outcome was the occurrence of at least one episode of infection during each posttransplant period. We excluded from the analysis asymptomatic bacteriuria, lower urinary tract infections (cystitis, urethritis and prostatitis), and upper respiratory tract infections not requiring hospital admission. The secondary study outcome was all-cause mortality. A number of pretransplant, perioperative and posttransplant variables were prospectively recorded. Glomerular filtration rate was estimated using the modification of diet in renal disease (MDRD-4) formula [13]. Since most of the plasma complement components are biosynthesized in the liver [9], we used serum albumin levels and internationalized normal ratios (INR) at months 1 and 6 as surrogate markers of liver function.

Immunosuppression and prophylaxis regimens

Detailed descriptions of the induction therapies and maintenance immunosuppressive regimens used in the cohort, as well as the antibiotic and anticytomegalovirus (anti-CMV) prophylaxis, are available in the Supporting Information that accompanies this paper.

Definitions

Hypocomplementemia was defined as a decrease in serum complement levels below the lower normal range given by our laboratory (as detailed above). Hypogammaglobulinemia (HGG) was defined as a serum IgG level <700 mg/dL [12]. The different infectious syndromes and acute graft rejection (AR) were defined according to commonly accepted criteria [14-18]. Detailed definitions of the outcomes analyzed in the study may be found in the Supporting Information.

Statistical analysis

Quantitative data were shown as the mean ± standard deviation (SD) or the median with interquartile ranges (Q1–Q3). Qualitative variables were expressed as absolute and relative frequencies with 95% confidence intervals (CIs). Categorical variables were compared using the χ2 test, Fisher´s exact test and McNemar test for repeated measures, as appropriate. Student's t-test or Mann–Whitney U test were applied for continuous variables. The linear associations between the normally distributed variables were assessed by Pearson's correlation coefficients. Patient and infection-free survival curves were estimated by the Kaplan–Meier method, and differences between groups were compared with the log-rank test. Backward stepwise Cox proportional hazards models were used to evaluate the association between hypocomplementemia and both primary ­and secondary study outcomes. Serum complement levels were also introduced into the models as continuous variables. Results are expressed as hazard ratios (HRs) with 95% CIs. Logistic regression was used to identify predictors of hypocomplementemia at baseline. Those variables found to be significant (p < 0.05) in the univariate analysis were included into the model in a backward stepwise fashion. We assessed the goodness of fit of the model using the Hosmer and Lemeshow test. Finally, predictors of hypocomplementemia during the posttransplant period were identified by mixed-effects logistic regression model, in order to take into account the within-subject correlation at different time points (months 1 and 6). The intraclass correlation (ICC) statistic represents the degree of association of the longitudinal data within subject, and specifically indicates the proportion of variance in the data attributable to individuals. A large ICC indicates that there is a large degree of homogeneity within subjects [19]. Results are given as odds ratios (ORs) with 95% CIs. All the significance tests were two-tailed. Statistical analysis was performed using SPSS v. 15.0 (Statistical Package for Social Sciences, Inc., Chicago, IL, USA), STATA v. 10 (Stata Statistical Software, College Station, TX, USA) and EPIDAT v. 3.1 (Conselleria de Sanidade, Xunta de Galicia, Spain).

Results

Baseline characteristics

We included 270 patients undergoing KT between November 2008 and May 2011 whose main clinical characteristics and posttransplant outcomes are summarized in Table 1. Assessment of serum complement factors at baseline was performed in all the participants. Samples at months 1 and 6 were available in 240 and 189 patients, respectively. We found no significant differences with regard to posttransplant infection or outcome between patients with immune assessment at months 1 and 6, and those from whom no specimens could be obtained (data not shown).

Table 1. Clinical characteristics and posttransplant outcome of 270 patients analyzed
VariableN(%)
  1. AR = acute graft rejection; CMV = cytomegalovirus; ESRD = end-stage renal disease; HLA = human leukocyte antigen; SD = standard deviation.

  2. 1Within the first posttransplant month.

Age of recipient, years (mean ± SD)54.6 ± 14.4 
Gender (male)170(63.0)
Pretransplant comorbidities  
Diabetes mellitus70(25.9)
Heart disease66(24.4)
Chronic lung disease36(13.3)
Chronic liver disease15(5.6)
Previous solid organ transplantation63(23.3)
≥2 previous transplants15(5.6)
Etiology of underlying ESRD  
Glomerulonephritis58(21.5)
Diabetic nephropaty49(18.1)
Nephroangiosclerosis35(13.0)
Policystosis34(12.6)
Chronic interstitial nephropathy23(8.5)
Unknown20(7.4)
Other51(18.9)
Baseline serostatus
Hepatitis C virus28(10.4)
Hepatitis B virus3(1.1)
CMV status D+/R–22(8.1)
CMV status D–/R–3(1.1)
Pretransplant renal replacement therapy
Hemodialysis221(81.9)
Continuous ambulatory peritoneal dialysis33(12.2)
Age of donor, years (mean ± SD)52.4 ± 16.6 
Type of donor
Brain-dead donor183(67.8)
Donor after cardiac death77(28.5)
Living donor10(3.7)
Number of HLA mismatches [mean (Q1–Q3 range)]4 (3–5) 
Cold ischemia time, hours (mean ± SD)16.9 ± 6.7 
Induction therapy
None53(19.6)
Basiliximab89(33.0)
Antithymocyte globulin127(47.0)
Primary immunosuppression scheme  
Tacrolimus, mycophenolate mofetil, and steroids240(88.9)
Tacrolimus, azathioprine, and steroids28(10.4)
Posttransplant complications  
Delayed graft function165(61.1)
Requirement of surgical intervention136(13.3)
At least one episode of AR63(23.3)
≥2 episodes of AR9(3.3)
Overall patient mortality17(6.3)
Infection-related mortality10(3.7)
Graft loss18(6.7)

Posttransplant outcomes

The median follow-up was 493.5 days (Q1–Q3 range, 404–738.7 days), with 231 patients (85.5%) reaching at least 12 months (total follow-up of 147 720 transplant-days). All-cause mortality was 6.3% (17 patients died at a median interval from transplantation of 174 days [Q1–Q3 range, 65–424 days]). One- and two-year survival rates were 95% and 92%, respectively. Death-censored graft survival at the end of follow-up was 92.9%.

Sixty-three patients (23.3%) had at least one AR episode during the follow-up period (mean, 1.14 ± 0.35 episodes per patient), including 11 cases of increase of serum creatinine without histological confirmation. The median interval to the first episode was 22 days (Q1–Q3 range, 12–77.25 days). The treatment for AR was based on the administration of steroid boluses (57 episodes), polyclonal immunoglobulins (32 episodes), plasmapheresis (20 episodes), rituximab (10 episodes) and antithymocyte globulin (5 episodes), alone or in combination.

A total of 147 recipients (54.4%) developed 320 episodes of posttransplant infection (overall incidence rate of 2.16 per 1000 transplant-days). The distribution of clinical syndromes and microorganisms is shown in Table 2.

Table 2. Site and causative agents involved in the 320 episodes of posttransplant infection
 N(%)
  1. APN = acute pyelonephritis; CMV = cytomegalovirus; HSV = herpes simplex virus; LRTI = lower respiratory tract infection.

  2. 1The total number may be less than the sum of each syndrome because more than one infection was simultaneously present in some patients (i.e. bloodstream infection associated to focal infection).

Clinical syndrome1  
APN54(16.9)
CMV viral syndrome74(23.1)
Incisional surgical site infection40(12.5)
Pneumonia and LRTI34(10.6)
Skin and soft-tissue infection23(7.2)
Digestive tract infection36(11.3)
Intraabdominal infection14(4.4)
Bloodstream infection64(20.0)
Infective endocarditis3(0.9)
Other11(3.4)
Isolated microorganisms  
Bacteria
Enterococcus spp.16(5.0)
Coagulase-negative staphylococci15(4.7)
Staphylococcus aureus5(1.6)
Viridans group streptococci2(0.6)
Escherichia coli44(13.7)
Klebsiella spp.7(2.2)
Other Enterobacteriaceae8(2.5)
Pseudomonas spp.21(6.6)
Other nonfermenting gram negative bacilli2(0.6)
Clostridium difficile14(4.4)
Mycobacterium tuberculosis complex2(0.6)
Polymicrobial5(1.6)
Other2(0.6)
No microbiological diagnosis30(9.4)
Viruses
CMV93(29.1)
HSV-1 and 211(3.4)
Varicella-zoster virus7(2.2)
Other11(3.4)
Fungi
Candida spp.14(4.4)
Aspergillus fumigatus2(0.6)
Other1(0.3)
Parasites (Leishmania donovani complex)3(0.9)

Timing and impact of hypocomplementemia on outcomes

As shown in Figure 1, the prevalence of C3 hypocomplementemia at baseline and month 1 was 21.5% (95% CI: 17–26%) and 20.4% (95% CI: 15–26%), respectively (p = 0.89), and decreased at month 6 to 11.6% (95% CI: 7–16%) (p = 0.017 compared to baseline). On the contrary, the prevalence of C4 hypocomplementemia significantly rose from 3.7% (95% CI: 1–6%) at baseline to 9.2% (95% CI: 5–13%) at month 1 (p = 0.004) and remained stable at month 6 (7.9%; 95% CI: 4–12%). Only four and five patients exhibited persistently low serum C3 and C4 levels throughout the entire study period, respectively. Figure 2 shows serum complement levels in patients with and without hypocomplementemia at different study time points. Pearson's correlation matrixes revealed a significant positive correlation for both C3 and C4 values at different time points (data available as Supporting Information). We also found a positive correlation between C3 and C4 levels at baseline (r = 0.441; p < 0.001) and months 1 (r = 0.450; p < 0.001) and 6 (r = 0.444; p < 0.001).

Figure 1.

Prevalence of C3 and C4 hypocomplementemia throughout the first 6 months after KT (C3 levels: p = 0.015 for the difference between baseline and month 6; C4 levels: p = 0.004 for the difference between baseline and month 1).

Figure 2.

Box-whisker plots showing serum complement levels in patients with (blank boxes) and without hypocomplementemia (gray boxes): C3 levels (2A) and C4 levels (2B) [median (dot), interquartile range (box), maximum and minimum values (whiskers), outlier values (circles) and extreme values (asterisks)].

Patients with either C3 or C4 hypocomplementemia at baseline did not experience higher rates of infection in the early period (first month) when compared to those with normal pretransplant levels. Recipients with C4 hypocomplementemia at month 1 only exhibited a higher cumulative incidence of acute pyelonephritis (p = 0.046) during the intermediate period, with no differences in other types of infection. Finally, there were neither significant differences in the incidence of infection during the late period in patients with or without C4 hypocomplementemia at month 6 (data not shown).

The incidences of infection during the posttransplant periods according to serum C3 levels are shown in Table 3. Patients with C3 hypocomplementemia at month 1 had a higher cumulative incidence of overall infection (p = 0.002), bacterial infection (p = 0.004), pneumonia (p = 0.043), fungal infection (p = 0.019), and non-CMV viral infection (p = 0.017) during the intermediate period (months 1–6). We also found a trend toward an increased cumulative incidence of CMV disease (p = 0.065). The incidence rate of overall infection and the mean number of infectious episodes per recipient were also higher in these patients as compared to those with normal C3 levels at month 1 (p < 0.001 for both). The presence of C3 hypocomplementemia at month 6 was associated with a higher cumulative incidence of bacterial infection in the late period (>6 month), although the difference did not attain statistical significance (p = 0.077). Figures 3 and 4 show overall infection-free survival during the intermediate period and bacterial infection-survival during the late period, respectively.

Table 3. Incidence of posttransplant infection according to the presence or absence of C3 hypocomplementemia
 C3 hypocomplementemia at month 1
Incidence of infection in the intermediate posttransplant period (months 1–6) (%)Present (n = 49)Absent (n = 191)p-Value
  1. APN = acute pyelonephritis; CMV = cytomegalovirus; HSV = herpes simplex virus; SD = standard deviation; VVZ = varicella-zoster virus.

  2. 1Data presented as number of infectious episodes per 1000 transplant-days.

Any infection28 (57.1)63 (33.0)0.002
Episodes (mean number ± SD)1.2 ± 1.30.6 ± 0.90.000
Incidence rate18.24.00.000
Bacterial infection18 (36.7)34 (17.8)0.004
Bloodstream infection6 (12.2)13 (6.8)0.167
Pneumonia6 (12.2)8 (4.2)0.043
APN6 (12.2)10 (5.2)0.082
Intraabdominal infection3 (6.1)6 (3.1)0.271
Fungal infection6 (12.2)6 (3.1)0.019
CMV disease17 (34.7)42 (22.0)0.065
CMV end-organ disease4 (8.2)8 (4.2)0.212
Non-CMV viral infection7 (14.3)8 (4.2)0.017
HSV or VVZ infection6 (12.2)3 (1.6)0.003
 C3 hypocomplementemia at month 6
Incidence of infection in the posttransplant period (>6 month) (%)Present (n = 22)Absent (n = 167)p-Value
Any infection6 (27.3)26 (15.6)0.142
Episodes (mean number ± SD)0.3 ± 0.70.2 ± 0.70.377
Incidence rate10.930.590.373
Bacterial infection5 (22.7)16 (9.6)0.077
Bloodstream infection2 (9.1)5 (3.0)0.190
Pneumonia2 (9.1)5 (3.0)0.190
Acute pyelonephritis1 (4.5)8 (4.8)0.718
CMV disease3 (13.6)12 (7.2)0.246
Non-CMV viral infection1 (4.5)8 (4.8)0.718
Figure 3.

Kaplan–Meier curves of infection-free survival throughout the intermediate period (months 1–6) according to the presence of C3 hypocomplementemia at month 1 (log-rank test; p = 0.001).

Figure 4.

Kaplan–Meier curves of bacterial infection-free survival throughout the late period (>6 months) according to the presence of C3 hypocomplementemia at month 6 (log-rank test; p = 0.017).

When focused specifically on the 31 patients (12.9%) with “de novo” C3 hypocomplementemia at month 1 (thus with normal levels at baseline), we found a higher cumulative incidence of bacterial infection during the intermediate period in this subgroup as compared with those with persistent hypocomplementemia (at baseline and month 1) and those with persistently normal C3 levels at both time points (41.9% vs. 27.8% vs. 15.8%, respectively; p = 0.003).

The Table 4 summarizes the results of multivariate Cox proportional hazards models for posttransplant infection. The presence of C3 hypocomplementemia at month 1 was identified as an independent risk factor for overall infection (HR 1.911; 95% CI: 1.175–3.108; p = 0.009) and bacterial infection (HR 2.130; 95% CI: 1.167–3.899; p = 0.014) during the intermediate period. C3 hypocomplementemia at month 6 predicted the occurrence of bacterial infection (HR 3.347; 95% CI: 1.061–10.562; p = 0.039) in the late period. When complement levels were analyzed as continuous variables, serum C3 levels at month 1 continued to exert an independent—and protective–effect on the incidence of overall (HR per unitary increment, 0.988; 95% CI: 0.979–0.997; p = 0.007) and bacterial infection (HR per unitary increment, 0.980; 95% CI: 0.968–0.992; p = 0.001) during the intermediate period.

Table 4. Multivariate analysis of risk factors for posttransplant infection (Cox proportional hazards model)
 Adjusted HR95% CIp-Value
  1. AR = acute graft rejection; CI = confidence interval; ESRD = end-stage renal disease; HGG = hypogammaglobulinemia; HLA = human leukocyte antigen; HR = hazard ratio; SOT = solid organ transplantation.

  2. *Per unitary increment.

  3. 1Model adjusted for number of pretransplant chronic comorbidities, glomerulonephritis as underlying ESRD, number of HLA mismatches, basiliximab as induction therapy, and requirement for reoperation within the first month.

  4. 2Model adjusted for number of pretransplant chronic comorbidities, diabetes mellitus as pretransplant condition, requirement for reoperation within the first month, occurrence of AR within the first month, and posttransplant renal artery stenosis.

  5. 3Model adjusted for recipient age, peripheral vascular disease as pretransplant condition, administration of steroid boluses within the intermediate period (months 1–6), serum albumin levels at month 6, IgG HGG at month 6, C3 hypocomplementemia at month 6, and graft function (MDRD-4) at month 6.

  6. 4Model adjusted for recipient age, number of pretransplant chronic comorbidities, peripheral vascular disease as pretransplant condition, glomerulonephritis as underlying ESRD, occurrence of AR within the first month, serum albumin levels at month 6, IgG HGG at month 6, and graft function (MDRD-4) at month 6.

Overall infection in the intermediate period (months 1–6)1   
Recipient age, years*1.0321.014–1.050<0.001
Posttransplant renal artery stenosis2.0171.168–3.4810.012
Serum albumin levels at month 1, g/dL*0.5420.361–0.8140.003
C3 hypocomplementemia at month 11.9111.175–3.1080.009
Bacterial infection in the intermediate period (months 1–6)2
Recipient age, years*1.0331.008–1.0580.008
Serum albumin levels at month 1, g/dL*0.3710.229–0.602<0.001
C3 hypocomplementemia at month 12.1301.167–3.8890.014
Overall infection in the late period (>6 months)3
Previous SOT3.1181.438–6.7600.004
Graft function at month 6, mL/min*0.9730.951–0.9950.018
Bacterial infection in the late period (>6 months)4
C3 hypocomplementemia at month 63.3471.061–10.5620.039

After excluding the only patient who died within the first month (day 14), we found that those recipients with C3 hypocomplementemia at month 1 had worse 2-year survival compared to those with normal values (84% vs. 94%; p = 0.029; Figure 5). Infection accounted for most of deaths in both groups (83.3% and 62.5%, respectively). Due to the low number of deaths, we could not perform a unique Cox regression model with all the potential risk factors for all-cause mortality. We alternatively attempted an exploratory approach by performing different Cox models that incorporated a maximum of two variables at a time—including C3 hypocomplementemia at month 1. After adjusting in pairs for recipient age, number of pretransplant chronic comorbidities, serum albumin levels at month 1, requirement for reoperation within the first month, and occurrence of delayed graft function, the presence of C3 hypocomplementemia remained as an independent risk factor for all-cause mortality, or showed a near-significant association with this outcome (minimum HR 2.709; 95% CI: 0.932–7.870; p = 0.067; remaining data not shown).

Figure 5.

Kaplan–Meier curves of patient survival according to the presence of C3 hypocomplementemia at month 1 (log-rank test; p = 0.029).

Risk factors for posttransplant hypocomplementemia

We further analyzed the risk factors for the development of C3 hypocomplementemia. We did not find any correlation between markers of liver function—serum albumin levels and INR—and complement levels at months 1 or 6, as assessed by Pearson's correlation coefficients. With regards to humoral immunity, patients with IgG HGG showed a lower prevalence of C3 hypocomplementemia at month 6 (2.4% vs. 13.7%; p = 0.031), with no differences at other time points (data available as Supporting Information). Only recipient age was associated with C3 hypocomplementemia at baseline in the multivariate analysis. The mixed-effects logistic regression model identified C3 hypocomplementemia at baseline, graft function at month 1, and the time interval elapsed from transplantation as predictors of the posttransplant occurrence of C3 hypocomplementemia throughout the first 6 months after transplantation (Table 5).

Table 5. Risk factors for the occurrence of C3 hypocomplementemia at baseline (binary logistic regression model) and throughout the first 6 months after kidney transplantation (mixed-effects logistic regression model)
 Univariate analysisMultivariate analysis
 OR95% CIp-ValueOR95% CIp-Value
  1. AR = acute graft rejection; CI = confidence interval; HGG = hypogammaglobulinemia; SOT = solid organ transplantation; OR = odds ratio.

  2. *Per unitary increment.

  3. 1Hosmer and Lemeshow test: p value = 0.428.

  4. 2Interclass correlation statistic: 0.390.

C3 hypocomplementemia at baseline1
Recipient age, years*0.970.95–0.990.0140.980.95–0.990.036
Diabetic nephropathy0.360.14–0.950.034 
C3 hypocomplementemia during the posttransplant period2
Pretransplant chronic lung disease2.341.04–5.270.035 
Pretransplant chronic liver disease3.130.77–12.850.122 
Pretransplant periodic hemodialysis4.321.28–14.590.011 
Previous SOT4.411.21–16.130.037 
Antihepatitis C virus positive status5.041.85–13.750.003   
C3 hypocomplementemia at baseline3.241.61–6.550.0016.802.42–19.12<0.001
Induction therapy2.400.89–6.460.074 
AR within the first month2.621.19–5.800.014-- 
Administration of steroid boluses3.480.99–12.270.063 
Graft function at month 1, mL/min*0.960.94–0.98<0.0010.970.94–0.990.005
IgG HGG at month 60.790.70–0.910.031 
Time interval since transplant      
Baseline to month 11 
Months 1–60.400.19–0.820.0120.410.20–0.830.014

Discussion

The risk of infection linked to over-immunosuppression remains as one of the major concerns faced by clinicians involved in the contemporary management of SOT recipients. Therefore, the search for immunological markers easily available at the bedside has become critical [20]. We found that the presence of C3 hypocomplementemia at month 1 after KT is associated with a twofold increase in the risk of overall and bacterial infection in the intermediate period (months 1–6) after adjusting for potential confounders—including serum albumin levels as surrogate marker of nutritional status and liver function. Analogously, C3 hypocomplementemia at month 6 increases by 3.3-fold the risk of late bacterial infection. To our knowledge, the present is the first study addressing the clinical value of monitoring of serum complement factors to identify KT recipients at highest risk of posttransplant infection. Moreover, our findings are in line with those reported in other types of SOT [10, 11].

End-stage renal disease (ESRD) due to diabetic nephropathy and recipient age were the only factors independently related to C3 hypocomplementemia at baseline in our cohort—exerting a protective effect in both cases. A population-based cohort study demonstrated that C3 levels were higher in men who subsequently developed diabetes, probably reflecting a low-grade systemic inflammation [21]. Serum C3 levels have been also associated with waist adiposity and insulin resistance, as adipose tissue may act as an alternative site of synthesis of this component [22]. Therefore, diabetic patients with ESRD seem to be protected to some degree against the development of hypocomplementemia.

There is scarce information regarding the kinetics of complement levels after KT [23-25]. Barnes et al. did not find significant differences between pre- and posttransplant values of C3 [23]. Accordingly, the prevalence of C3 hypocomplementemia was similar at baseline and at month 1 in the present cohort (≈20%), and only fell in the late period. A recent study reported a prevalence of C3 hypocomplementemia very close to ours (19%) in stable KT recipients [25]. The presence of low complement levels already at baseline emerged as the strongest risk factor for the occurrence of C3 hypocomplementemia throughout the first 6 months. The incorporation into the Banff criteria of C4d staining in the peritubular capillaries exemplifies the role of the classical activation pathway in the antibody-mediated alloreactivity [26]. The occurrence of AR has been associated with a decrease in serum complement levels [23, 25]. We found that AR within the first month increased the risk of C3 hypocomplementemia during the subsequent period. Interestingly, graft function at month 1 emerged as an independent predictor of C3 hypocomplementemia, maybe reflecting the deleterious impact of early AR episodes on the serum complement levels. Therefore, we hypothesize that low complement levels are plausibly secondary to rejection-associated complement activation, rather than to a defective synthesis. Supporting this, we did not find any correlation between markers of liver function and complement levels at months 1 or 6. Unfortunately, the lack of measurements of complement activation hinders further insight into the mechanisms leading to hypocomplementemia.

The importance of the innate immunity—and specifically complement system—is increasingly being recognized in situations of impaired adaptive immunity such as immunosuppression following SOT [27]. Hypocomplementemia at baseline did not have any significant effect on the risk of early infection in our study. This finding may be justified by the predominance of donor-derived and device-associated nosocomial infections during the first posttransplant month, with a minor impact of recipient's net immunosuppressive state [20]. On the contrary, the presence of C3 hypocomplementemia beyond the first month exerted a significant impact on the incidence of overall and bacterial infection, as well as on some specific infectious syndromes.

Cumulative incidence of pneumonia was threefold higher in those recipients with lower C3 levels, with the difference achieving statistical significance during the intermediate period. Experimental models of pneumonia have highlighted the role of C3 in pulmonary defence against Streptococcus pneumoniae and Pseudomonas aeruginosa [28, 29], further supported by previous clinical observations [30]. Both CMV and other viral infections were also more frequent in patients with hypocomplementemia at month 1. C3 has been shown to be critical in B cell activation and T cell dependent antibody responses during viral infections, as demonstrated for herpes simplex [31] or influenza viruses [32]. MBL deficiency has been identified as a risk factor for the development of CMV infection in KT recipients in some studies [6], but not in others [8]. Finally, the cumulative incidence of fungal infection during the intermediate period, mainly due to Candida spp., was higher among hypocomplementemic patients. It has been reported that C3-deficient mice are exposed to an increased susceptibility to Candida infection and delayed fungal clearance [33]. Sarmiento et al. found no association between C3 levels and the incidence of fungal infection—mainly invasive aspergillosis—in heart transplant recipients [11].

Despite the pivotal role of C4 in the classical activation pathway of the complement system, we found no evidence that serum levels of this component influence the risk of infection. This observation is supported by that previously reported by Wahrmann et al. showing no significant effect of C4 gene copy-number variations—the number of inherited copies closely correlates with serum C4 concentrations and complement activity—on kidney allograft outcome or infection-related hospitalization rate [34].

Interestingly, C3 hypocomplementemia was associated with a worse patient survival in the present study. On the opposite, Berger et al. demonstrated superior graft and patient survival in KT [35] and pancreas–kidney recipients [36] with low MBL levels. One explanation for these apparently conflicting findings might lie in the fact that most of graft losses and deaths in these later studies were due to treatment-resistant AR and cardiovascular complications, whereas most of deaths in our cohort were infection-related. Therefore, complement-status seems to be a two-sided sword in posttransplant outcome, with different impacts on the risk of infection, ischemia-reperfusion damage, and cardiovascular events.

Despite its notable strengths—large sample size, prospective design with scheduled sampling times, accurate assessment of the type of infection in each posttransplant period, and relatively low loss-to-follow-up rate—the present study also has some limitations. The single center nature of our findings hinders its generalizability to other institutions with different posttransplant strategies in prophylaxis or routine laboratory testing. We did not perform functional complement assays (i.e. CH50) or measure cleavage products of C3 or C4 (i.e. C3d or C4d). As commented, such limitation prevented us from assessing whether hypocomplementemia was the result of complement activation, impaired synthesis, or increased consumption. In addition, we did not compare the prognostic impact of C3 hypocomplementemia and MBL-deficient states, whose predictive value for posttransplant infection is supported by an increasing amount of evidence [2-8]. Our study also lacks measurement of C-reactive protein (CPR) levels, the comparison of which with serum complement would have been of interest. Nevertheless, the inclusion of serum albumin as inverse acute-phase reactant in all the multivariate models could partially overcome this limitation, since albumin levels exhibits a strong and negative correlation with CPR in hemodialysis patients [37]. Finally, the sample size might not be large enough to detect differences in incidence rates of late infection.

In conclusion, this study suggests that a simple monitoring schedule of serum C3 levels during the posttransplant follow-up can effectively predict the risk of infectious complications in KT recipients. The measurement of serum complement levels by nephelometry is an affordable and broadly available technique, thus contributing to the eventual cost-effectiveness of such strategy. Further studies are needed to enlighten the role of complement system in the pathogenesis of infection after KT, as well as its potential clinical implications. Meanwhile, our findings may open new perspectives in the design of individualized approaches for posttransplant prophylaxis, tapering of immunosuppression, and preemptive therapy.

Acknowledgments

Funding sources: This study was supported by the Spanish Ministry of Economy and Competitiveness (Fondo de Investigaciones Sanitarias [FIS] project 11/01538) and the Fundación Mutua Madrileña de Investigación Médica (FMM Grant 2010/0015). Mario Fernández-Ruiz holds a research training contract “Rio Hortega” (CM11/00187) from the Spanish Ministry of Economy and Competitiveness (Instituto de Salud Carlos III). Francisco López-Medrano is currently receiving a grant from the Fundación Mutua Madrileña.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary