World Journal of Urology

, 25:445

Renal transplant rejection markers

Authors

    • Department of Internal Medicine, Division of NephrologyMedical School Hannover
Topic Paper

DOI: 10.1007/s00345-007-0211-6

Cite this article as:
Gwinner, W. World J Urol (2007) 25: 445. doi:10.1007/s00345-007-0211-6

Abstract

Acute rejection is one of the key factors which determine long-term graft function and survival in renal transplant patients. Timely detection and treatment of rejection is therefore, an important goal in the post-transplant surveillance. The standard care with serum creatinine measurements and biopsy upon allograft dysfunction implies that acute rejection is detected in an advanced stage. Therefore, non-invasive monitoring for acute rejection by markers in blood and urine has been tried over the past decades. This review describes the requirements that should be met by non-invasive markers. The experience with single biomarkers and with newer approaches—mRNA expression analysis, metabolomics, and proteomics—will be discussed, including future directions of necessary research.

Keywords

KidneyTransplantationBiomarkerAcute rejectionProteomics

Rejection and renal allograft survival

In the past decades, progress has been made in the short-term success of renal transplantation which is mainly attributable to more efficient immunosuppressive therapies. However, long-term allograft survival has not improved similarly [1]. The leading cause for allograft loss after the first year of transplantation is a chronically-progressive scaring process of renal tissue [2], formerly called “chronic allograft nephropathy” [3]. The exact causes of this nephropathy are largely unknown. Acute rejection processes, even when subclinical (i.e., without concomitant allograft dysfunction) [4, 5], are an established, important factor. Other causes include acute tubular injury, toxic effects of drug treatments—especially from calcineurin inhibitors, viral graft infections and factors related to the graft quality (prolonged cold ischemia time, high donor age, graft origin from deceased donors) [6, 7]. Further, chronic rejection processes are implicated in the progressive scaring of renal allografts [3], yet, the understanding of this in terms of definition and pathophysiology is less clear than for acute rejection.

Because of the central role of acute rejection for the long-term allograft survival, timely diagnosis and therapy is a foremost task in the care of renal transplant patients. Reliable and early detection of rejection is particularly important in the efforts to spare patients from over-immunosuppression by using minimized and tailored immunosuppressive protocols. This article discusses current and future concepts of detecting acute rejection, with special emphasis on non-invasive approaches.

Heterogeneity of rejection processes

Before going into the tools for rejection diagnosis it must be realized that acute rejection is a heterogeneous condition with regards to the involved renal structures, the mechanisms of immunological injury, and the clinical presentation.

Major targets of immunological injury include tubular epithelial cells with their adjacent structures (peritubular capillaries, tubulointerstitium), glomerular capillaries and larger renal vessels. Categorization of this multiplicity has been sought by the BANFF classification [8], which is widely used for the evaluation of renal allograft biopsies. Besides topographical categorization of immune injury in this classification, the changes are rated by their severity. Altogether, these scores are summarized into a simple scheme of borderline rejection, and different severity grades of acute cellular tubulointerstitial and vascular rejection. In addition, a newly included entity is antibody mediated, humoral rejection, detected by the complement fragment C4d at the level of peritubular capillaries [9].

As denoted by the broad BANFF classes, different immune mechanisms cause acute rejection. Cellular rejection is characterized by T-cell infiltration, yet, inflammatory infiltrates may contain different subsets of T-cells, B-cells, histiocytes, macrophages and granulocytes [10]. Antibody-mediated rejection on the other hand, essentially involves B-cell populations. On a molecular level, interaction, activation and injury mechanisms of the different cell types is interrelated to cytokines, chemokines, adhesion molecules, complement factors, proteases and signaling molecules [11], which can be anticipated to exhibit variable expression in the known rejection types.

In clinical terms, heterogeneity of acute rejection is mainly characterized by severity and time-related aspects. Rejection severity may range from subclinical episodes (detected in protocol biopsies) up to full-blown acute renal failure. Time-related aspects refer to the dynamics of acute rejection (incipient or advanced stage) and to the time interval from transplantation until rejection. Particularly, late occurring acute rejections carry a worse prognosis for the allograft [6]. The use of different immunosuppressive drug combinations with agents that deplete T-cells, B-cells or monocytes/macrophages or block co-stimulatory pathways, calcineurin inhibitors and anti-proliferative drugs can be expected to add to the heterogeneity of the clinical, pathomechanical and morphological features [12] of acute rejection and has been shown to affect rejection markers [13].

Current approaches to diagnose rejection

Common post-transplant surveillance primarily relies on monitoring of serum creatinine levels, besides recognition of non-specific signs which may be present in acute rejection like reduced urine output, hypertension, graft tenderness and peripheral edema [14]. Serial ultrasound examination in the early post-transplant period is performed in some transplant centers to detect disturbances in the graft, yet, issues of specificity and sensitivity for acute rejection need to be considered, even with the use of echo enhancers [15]. An increase in serum creatinine is suspicious of rejection and some clinicians will proceed with anti-rejection treatment after excluding other causes (exsiccosis, urinary tract infection or obstruction, allograft artery stenosis, toxic drugs effects). Ideally, a biopsy is performed to confirm the presence of rejection and to determine its type, as this may translate into different therapeutic interventions [9].

Surveillance of patients by this approach implies recognition of the rejection at a relatively late stage, i.e., when immunological injury has reached a degree that causes impaired graft function. Therefore, some transplant centers have turned into regularly performed protocol biopsies to detect acute rejection earlier. Considerably high frequencies of acute rejection (mostly subclinical) have been reported in protocol biopsies, with up to 33% at 3 months post-transplantation [16]. An inherent limitation of this approach is that multiple biopsies are required to pick up the rejections as early and frequently as possible. Further, the sampling error (i.e., missed rejection diagnosis due to the focal nature of infiltrates and limited sample sizes), the individual variability in evaluating biopsies, the costs and procedural risks need to be considered [17, 18].

Principal expectations and requirements towards non-invasive markers for acute rejection (Table 1)

To compensate for the limitations of allograft biopsy, many attempts have been made to establish surrogate markers of rejection. Yet, none of these markers have made their way into today’s clinical practice as a universally accepted diagnostic tool suggesting that the ideal marker(s) is still to be discovered.
Table 1

Expectations towards non-invasive rejection markers

Sampling

Easy access to samples (e.g., blood, urine)

Repeated/serial sampling possible

Marker stability (with regards to storage period, temperature, pH, sample constituents like proteases)

Analysis

Simple and reproducible method

Time and cost efficiency of the test

Robustness against interfering substances of the sample

Validity for different sample dilutions (urine)

Explanatory power

Overall sensitivity and specificity for discriminating rejection from no rejection

Ability to discriminate time stages of rejection (incipient vs. advanced)

Ability to discriminate severity grades of rejection (mild vs. severe)

Ability to detect different rejection types (T-cell mediated tubulointerstitial/vascular rejection, antibody mediated rejection)

What is an ideal marker? Sample collection should be easy and repeatedly possible which is true for urine and blood. Stability over time (ideally at ambient temperature) and insensitivity against pH changes and other sample constituents (e.g., proteases) are important, which requires to consider sampling conditions (first voiding vs. mid-stream urine, 24-h collections). Analysis of the marker should be methodologically simple, reproducible, and time and cost efficient. Insensitivity to inherent interfering substances in the sample (e.g., high glucose levels in blood, protein and salts in urine) is desirable.

Sensitivity and specificity relate to the practical value of a marker i.e., to its reliability to detect rejection and to differentiate rejection from non-rejection. This is especially critical in mild or incipient rejection and in conditions with allograft infection (e.g., bacterial, polioma virus, CMV) which need to be distinguished from inflammation caused by rejection [19].

Another question is how differentiated markers describe the rejection. The simplest means would be to determine if acute rejection is present, based on the detection of a marker which is not existent in non-rejecting patients, or, on the absence of a marker normally seen in patients without rejection. Yet, in view of the heterogeneity and complexity of acute rejection, it is unlikely that one marker will fit all facets of this process. Thus, distinct markers may detect incipient, fully established and resolving stages of rejection. Very likely, different rejection types (T-cell vs. antibody-mediated, tubulointerstitial vs. vascular) have distinct sets of markers due to diverse pathomechanisms of each type as suggested by DNA array analyses of renal allograft biopsies [20]. Further, quantifiable markers could give additional information about the severity of rejection.

Concepts of detecting acute rejection by non-invasive markers

Use of individual immune and injury markers

Present approaches are based on two principles. One is to monitor aspects of the immune system which includes alloimmune recognition/activation and the effector pathway of inflammation. Table 2 shows a comprehensive list of individual markers which have been used so far [14, 2156]. Not included are analyses of HLA antibodies by flow-cytometry or ELISA and of alloreactive T cells by ELISPOT as these tests estimate rather the immunological risk, instead of pinpointing the acute rejection (for review see [57]). The other principle is to detect injury to tubular epithelial cells, the renal interstitium, and to vascular and glomerular structures. However, these tests (e.g., circulating endothelial cells, β2-microglobulin in serum and urine, serum uPAR, urine neutrophil gelatinase-associated lipocalin and interleukin-18) may be confounded by ischemia/reperfusion and toxic drug injury [5863].
Table 2

Single markers for acute rejection

 

Marker

Sensitivitya (%)

Specificitya (%)

Remarks

References

Peripheral blood cells

 Mononuclear cells

mRNA for: IL-4, IL-5,IL-6, IFN-γ, granzyme B, perforin

75

85

 

[21]

 Mononuclear cells

mRNA for: granzyme B, perforin, HLA-DRA

   

[22]

 Leukocytes

mRNA for: perforin, granzyme B, FasL

100

95

 

[23]

 CD4+ T cells

CD40L 

   

[24]

 CD3+ T cells,

 CD8+ T cells

CD69

   

[25]

 CD8+ lymphocytes

S6F1+

87

83

 

[26]

Blood/serum/plasma

 Blood

Perforin mRNA

Granzyme B mRNA

90

87

82

72

2–30 days positive before overt rejection

[27]

 Serum

Myeloid-related protein

67

100

With a median of 5 days positive before overt rejection

[28]

 Serum

IL-2R

   

[29]

 Serum

Neopterin, sIL-2R

  

Confounder: renal function, CMV infection

[14, 30]

 Plasma

sCD30

88

100

Positive before overt rejection

[31]

 Plasma

sCD44

   

[32]

 Plasma

TNF-α, TNF-β, sICAM

  

2–3 days positive before overt rejection

[33]

 Plasma

IL-2, IL-3, IL-4, IL-6, IL-8

   

[34]

 Plasma

e-selectin, ICAM-1,VCAM-1

   

[35]

Urine and urinary cells

 Urine

sVCAM-1,C4d

   

[36]

 Urine

Monokine induced by IFN-γ (MIG), IP-10

93

89

Positive before overt rejection in some cases

[37]

 Urine

IP-10, MIG, IFN-inducible TC chemaoattractant (CXCL11)

86

91

confounders: ATN, BK virus nephritis

[38]

 Urine cells

Granzyme B mRNA

Perforin mRNA

83

83

 

[39]

 Urine sediment

NNKG2D receptor mRNA

   

[40]

 Urine sediment

mRNA for: granulysin, granzyme B, perforin, FasL, TNF-α, RANTES, IL-2, IL-10, INF-γ,TGF-β, CD3, CCR1

  

granulysin positive before overt rejection

[41]

 Urine

IP-10 (CXCL10) mRNA and protein

  

Positive before overt rejection

[42]

 Urine

FOXP3 mRNA

90

73

Inversely correlated with S-creatinine

[43]

 Urine cells

HLA-molecule expression, IL-2R

   

[44]

 Urine cells

CD2, CD4, CD8, CD25, HLA-DR 

   

[45]

 Urine cells

Serine proteinase inhibitor PI-9

   

[46]

 Urine cells

Granzyme B mRNA

  

Urinary tract infection not confounding

[47]

 Urine cells

CD103 mRNA

59

75

 

[48]

 Urine cells

IP-10 mRNA

CXCR3 mRNA

100

63

78

83

 

[49]

 Urine

T-cell receptor subtype analysis

   

[50]

 Urine

sIL-2R 

  

Serum IL-2R level affects diagnosis

[51]

 Urine

IL-6

  

Confounders: infection, acute tubular necrosis

[52]

 Urine

IL-8

  

Positive before overt rejection confounders: acute tubular necrosis, infection

[53]

 Urine

MIF, macrophage migration inhibitory factor (MIF)

  

Positive before overt rejection

[54]

 Urine cells

HLA-DR, CD3

100

93

 

[55, 56]

aIf study examined several markers, single sensitivity and specificity measures refer to the most sensitive and best discriminating markers according to the evaluation of the study investigators

Blood-derived markers

Blood tests appear to be especially suitable to monitor the immune system. Yet, this relies on the assumption that immune processes towards the allograft are reflected by circulating leukocytes and their mediators. Blood could be also used to detect allograft injury as local factors from the graft might pass over into the blood circulation. This may be particularly applicable to endothelial injury, which occurs in vascular rejection, glomerulitis and peritubular capillaritis [59].

Urine-derived markers

Urine is widely used in the diagnosis of renal disease, however, the information from this source is quite complex. Markers derived from blood via glomerular filtration may be variably present, dependent on the glomerular filtration rate and permeability, and on the marker size, charge and blood concentration [30, 43, 51, 63]. Markers can originate from nephron structures by secretion, transcellular migration, and release or shedding from injured cells. Such local factors may particularly reflect rejection-associated dysfunction and injury of podocytes and tubular epithelial cells which are in close contact with the urine. Of note, the whole concept of filtrated and locally released markers has to consider alterations of these factors during tubular passage (e.g., reabsorption). Also, the entire urogenital tract contributes to the urine composition which is important when allograft or urinary tract infection is present [38, 52, 53].

Use of marker combinations

Limited sensitivity and specificity in the use of single markers has led to the concept of combining markers to improve the detection of rejection. This is not confined to the mathematical fact that combination of markers (with their individual sensitivities and specificities) increases the overall sensitivity and specificity. From a biological view, combined markers could reflect variations in the rejection process better. For example, markers for immune activation will appear early in the rejection process, like cytokine secretion from antigen presenting cells. Markers representative of effector pathways (e.g., granzyme B, perforin) or local graft injury (e.g., endothelial adhesion molecules, chemokines) can be expected in later stages [11]. Also, the rejection severity can result in heterogeneous marker expression, like the release of tubular enzymes in mild, rejection-induced tubular dysfunction and reduced tubular β2-microglobulin reabsorption in severe damage [60]. Complex rejections types with tubulointerstitial and vascular involvement are another example where the combined use of markers for vascular and tubular injury may be advantageous.

New approaches

Traditionally, selection and establishment of the markers has been based on their known role in rejection. A novel avenue to define rejection markers is characterized by a reverse approach that is represented by the technologies of mRNA expression profiling, proteomics and metabolomics. It uses the entire information of mRNA expression, proteins and peptides, or non-protein metabolites that is detectable in the sample. Compared to analyzing known, pre-defined rejection markers this is a relatively unbiased approach which is, in essence, only limited by the thresholds to detect and discriminate the variety of molecules of a sample. Applying bioinformatics, the entire molecular information is then utilized to establish sets of molecules which are representative of rejection compared to samples without rejection. A major challenge and pre-requisite of this approach is to exactly define the rejection phenotypes and to keep track of potentially confounding conditions like infectious, toxic, and ischemic injury. The potential of the new technologies is not confined to the discovery of useful rejection markers. Novel mediators of the rejection process may be identified which could lead to new therapeutic targets.

mRNA expression analysis (Fig. 1)

mRNA analysis with DNA microarrays has been used on renal allograft biopsies and peripheral blood lymphocytes. In biopsies, this has illustrated the enormous heterogeneity of acute rejection [20]. Among hundreds of differentially expressed genes, transcripts from T and B cells, natural killer cells and macrophages were found, reflecting the complex interaction of different leukocyte populations. Further, early mediators (e.g., interferon-γ, interleukin-2 receptor components) and late mediators downstream of T-cell activation (e.g., granzyme A, RANTES) were shown to be involved. Interestingly, a subset of biopsies that were histologically indistinguishable from other rejection samples, displayed an overriding signature for CD20+ B-cells which was clinically associated with steroid-resistant rejection. Additional characteristic signatures were identified for nephrotoxic drug effects, urinary tract infection, and chronic allograft nephropathy. With a similar approach [64], up-regulation of genes involved in inflammation and down-regulation of genes for basic cellular metabolism were found in biopsies with acute rejection. Distinct gene signatures separated acute rejection from other causes of allograft dysfunction. As a non-invasive approach, the study also illustrated the potential to differentiate rejecting from non-rejecting patients by mRNA analysis of peripheral blood lymphocytes.
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-007-0211-6/MediaObjects/345_2007_211_Fig1_HTML.gif
Fig. 1

mRNA expression analysis by DNA microarrays. Shown is an example of the differential analysis of samples with and without rejection. The mRNA is isolated from the cells of the samples, reverse transcribed into cDNA and tagged with different fluorescent labels to discern the rejection sample from the non-rejection sample. The cDNA is hybridized to the immobilized DNA templates of the array in which each spot with a certain template corresponds to a particular gene (DNA chip). The fluorescence of bound cDNA is detected and differential gene expression profiles are constructed from these data, identifying genes or gene clusters which are preferentially expressed in non-rejecting and rejecting samples

Metabolomics

This is an emerging branch of research that employs high-throughput measurement of large numbers of small (<1,500 Da) molecule metabolites of endogenous and exogenous origin. As with DNA microarray analysis and proteomics, metabolomics requires advanced separation techniques and bioinformatics to analyze the wealth of information present in blood or urine. Potential applications include assessment of ischemia-reperfusion injury and drug toxicity, its use in rejection diagnosis has not been evaluated so far [65].

Proteomics

A few studies have used proteomics in renal transplant patients. Proteomics examines the actual abundance of proteins and peptides. This is different to mRNA expression profiling which does not necessarily reflect protein expression due to the in vivo variability in mRNA stability, translation and post-translational modification.

The basic steps in proteome analysis (Fig. 2) include separation of proteins/peptides by two-dimensional electrophoresis (2-DE), surface-enhanced laser desorption ionization (SELDI), capillary electrophoresis (CE), or liquid chromatography (LC), followed by mass spectroscopy (MS) of the molecules. Bioinformatics is then used to translate the mass spectroscopy data into peptide lists and to discern peptide patterns that correlate with pre-defined phenotypes. Optionally, peptide sequencing can be performed afterwards [66]. It is beyond the scope of this article to discuss the different techniques of proteome analysis (2-DE–MS, SELDI–MS, LC–MS, and CE–MS; for details see [67]). In practice, studies in renal transplant patients have employed SELDI–MS, a relatively simple method with a lower peptide yield and resolution, and CE–MS, which has higher sensitivity and sample resolution.
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-007-0211-6/MediaObjects/345_2007_211_Fig2_HTML.gif
Fig. 2

Proteome analysis. After sample preparation by different cleaning steps the proteins and peptides are separated by two-dimensional electrophoresis, surface-enhanced laser desorption ionization, capillary electrophoresis (shown in this scheme), or liquid chromatography. Mass spectroscopy of the molecules is then performed, followed by translation of the mass spectroscopy data into peptide lists. Peptide patterns from samples with and without rejection are compared to define sets of peptides which can discriminate between rejection and non-rejecting states

Using SELDI–MS on urinary samples, 22 stable patients without evidence of acute rejection (biopsy-confirmed in 12 cases) were compared with 23 patients with biopsy-proven acute rejection (BANFF grade not reported). Rejecting patients were identified by the presence of 4–7 peptides, with a sensitivity of >90% and a specificity of approximately 80% [68].

In another study, SELDI–MS of urines from 17 patients with biopsy-proven acute rejection (BANFF grade not reported) and 15 non-rejecting patients showed 45 different protein peaks between the groups including some down-regulated peaks. Finally, when applied on the sample sets used to establish these protein patterns, two of the markers (10.0, 3.4 kDa) were sufficient to yield 83% sensitivity and 100% specificity for rejection diagnosis [69].

A third study examined 15 patients with acute tubulointerstitial rejection and 3 cases with vascular rejection in comparison to 22 patients who had not experienced delayed graft function or rejection and who were without chronic changes in a 12-months protocol biopsy. Three protein peaks were found in the urine of 17 from 18 patients with rejection but only in 4 samples from non-rejecting patients by SELDI–MS. Some patients had CMV viremia but apparently, this did not change the peptide pattern in both groups [70]. In a follow-up study, the protein peaks were identified as cleavage products of β2-microglobulin [71].

In our own study [72], we used CE–MS to establish urine peptide patterns from patients with acute tubulointerstitial rejection and vascular rejection (subclinical and clinical), and compared these to non-rejecting cases and to patients with urinary tract infection. Patients had variable concomitant pathological findings in the biopsies, like tubulointerstitial fibrosis, acute tubular injury and calcification. Also, clinical factors were quite variable (diabetes mellitus, smoking, CMV viremia, hematuria, immunosuppressive regimens, other drug therapies). The established patterns were validated with blinded samples. Two of ten control samples without acute rejection and urinary tract infection were misclassified. Of the nine samples with acute rejection, one was classified as having no rejection. In addition, two other samples which revealed only borderline rejection on re-review of the biopsies, were not classified as rejection.

In summary, these studies have principally demonstrated the feasibility of mRNA expression analysis and proteomics to detect acute rejection. Compared to the use of single or few markers, these new approaches may better account for the heterogeneity of rejection processes, resulting into refined sets of markers for specific types and clinical features of acute rejection.

Practical application of non-invasive markers

A substantial restriction must be addressed in the first place: none of the markers or marker sets has been validated to an extent which allows well-founded recommendations for their use. Most studies on rejection markers are small, allograft biopsies are not always available for correlation, and potentially confounding, co-existing conditions are rarely appreciated completely. Nevertheless, based on the frequency and agreement of previous study results, combined measurement of granzyme B and perforin mRNA in blood or urine samples and urinary IP-10 (interferon-γ inducible protein; CXCL10) mRNA or protein promise a relatively sensitive and specific detection of acute rejection, at least for the most common, tubulointerstitial type. Potentially, subclinical rejection is also captured by this combination as positivity of these markers prior to overt clinical rejection has been reported (Table 2). Alternatively, mRNA expression analysis or proteomics afford a similar high sensitivity and specificity but may account better for the broad spectrum of rejection as discussed above. Yet, effectiveness of the different tests in terms of efficacy versus expenditure of time and costs has not been evaluated so far.

In practice, monitoring with combinations of granzyme B, perforin and IP-10 could be regularly done in stable patients, particularly in the first year after transplantation where acute rejections occur most often. Alternative tests like proteome analysis may substitute for the mentioned markers once superiority of these tests has been confirmed by comparative studies. If one of the non-invasive tests suggests acute rejection, an allograft biopsy is indicated. In patients with graft dysfunction, a biopsy is indicated even if non-invasive markers are negative because detection of cellular vascular and antibody mediated rejection types has not been sufficiently verified for the non-invasive tests (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-007-0211-6/MediaObjects/345_2007_211_Fig3_HTML.gif
Fig. 3

Algorithm for the use of non-invasive markers. Monitoring by regular serum creatinine measurements should be paralleled by examination of non-invasive rejection markers. In stable patients (unchanged serum creatinine) detection of IP-10 which is involved in immune activation and of the cytotoxic effector molecules granzyme B and perforin [11] suggests acute rejection. An allograft biopsy should be performed to clarify if rejection is present. As non-invasive markers have not been validated to detect vascular or antibody mediated rejection, the biopsy is also important to detect concomitance of these rejection types. For the same reason, patients with an increase in serum creatinine should have a biopsy after excluding other obvious causes for graft dysfunction, even if non-invasive markers do not suggest acute rejection

Summary and future directions

It is clearly desirable to have non-invasive markers to monitor renal transplant patients closely for acute rejection. Yet, despite the long-standing interest and efforts to establish markers for acute rejection, post-transplant surveillance still relies mostly on creatinine measurement and allograft biopsy. To overcome the limitations which have impeded non-invasive markers to become a generally accepted tool for the clinical practice, further work is necessary.
  1. 1.

    Establishment of markers should be strictly based on the histologically defined entities and severity grades of acute rejection.

     
  2. 2.

    The clinical picture of acute rejection (early vs. late stages, subclinical vs. clinical rejection) has to be incorporated in the establishment of markers.

     
  3. 3.

    Confounding factors and conditions—either present in the biopsy or clinically existing—should be systematically considered during establishment of markers.

     
  4. 4.

    Standardization of sample collection and processing and of analysis procedures is required.

     
  5. 5.

    Sufficiently large prospective studies with blinded samples are required to confirm the usability of the markers and to determine their actual sensitivity and specificity.

     
  6. 6.

    In the process of validation, the limitations of markers have to be defined like confounding factors and the thresholds of detection in terms of severity and time-dependent aspects of acute rejection.

     
  7. 7.

    Eventually, for advanced techniques like proteomics it remains to be explored if sets of markers can be identified that are suitable for simplified detection techniques.

     

Copyright information

© Springer-Verlag 2007