Introduction

Biopsies of kidneys from deceased donors are most commonly taken to assess quality of the parenchyma, not of tumours. Precise assessment of the quality of donor organs is a prerequisite for fair allocation and maximal exploitation of the donor organ pool. Donor kidneys classified as “marginal” solely on clinical grounds can have a wide variability of histological findings [1] and can perform quite well, if carefully selected [2]. It is our strong opinion that histopathologic examination can significantly contribute to successfully tapping the pool of donor kidneys deemed “marginal.” Moreover, novel therapeutic options to mend the quality of donor organs with perfusion systems are gaining attention [3]. Although in most cases, the donor kidney will already be on a perfusion machine when the results from tissue diagnostics are available, it could still direct the use of special pharmacologic agents in the perfusate.

Clearly, the decision about the biopsy of a donor kidney should remain the privilege of the allocating clinician. Histopathologists are only involved, when the biopsy has been taken. Since at present, zero-hour biopsies are more exception than the rule in German practice, kidney biopsies from deceased donors are subject to bias toward the so-called marginal organs.

Only in very few cases histopathologic findings alone can preclude transplantation of a donor organ. In the vast majority of cases, also clinical findings from both donor and recipient have to be carefully taken into account. The urge to give a transplant to a patient suffering on dialysis has to be weighed with the inherent risk of kidney transplantation: Delayed graft function, the risk factors of which are mostly related to the donor and to ischemia/reperfusion injury, endangers the life of the donor organ as well as the life of the recipient [4–6]. Again, histopathology should try to offer help with these difficult decisions.

Development of the recommendations

The following recommendations for the biopsy procedure itself, the laboratory workup and the histopathologic evaluation, for the communication and collection of results have been developed during two joint workshops held by the German Organ Transplantation Foundation (Deutsche Stiftung Organtransplantation - DSO) and the German Society of Pathology (Deutsche Gesellschaft für Pathologie - DGP) in Hannover on the 10th of November 2010 and in Frankfurt/Main, Germany on the 23rd of October 2013. JUB has been commissioned by the last joint workshop in 2013 to write these recommendations. PP, IH and JUB have been commissioned by the assembly of the Kidney Committee of the Deutsche Transplantationsgesellschaft (DTG) on the 26th of October 2013 to write and publish these recommendations. The kidney committee has accepted these recommendations at the assembly on the 16th of October 2014. Thus, these recommendations are supported by the DSO, the DGP and DTG.

These recommendations are based on the published studies listed in Table 1 and in the reference section, and discussions at the workshops and meetings mentioned above and on unpublished experiences of the authors. Most of the studies are retrospective studies with a bias toward marginal donors. While some of the scores and indices are based solely on histopathologic findings [1, 7, 8], others also incorporate clinical findings [9, 10]. The most decisive parameters according to the literature seem to be arterial intimal fibrosis, the percentage of globally sclerosed glomeruli and the extent of cortical tubular atrophy and interstitial fibrosis. However, none of these scores have gained widespread acceptance. Some might even be unsuitable for German practice. For example, the authors of the Maryland Aggregate Pathology Index (MAPI) claimed that their score needed validation for the kidneys with a cold ischemia time (CIT) of less than 24 h [1], which the vast majority of renal transplants in Germany have. Thus, our recommendations at this stage are only a standardised semiquantitative description of histopathologic findings. Based on this standardisation of histopathological findings and the biopsy procedure itself, we will aim to build a database for prospective studies in Germany under the tutelage of the DSO (see discussion and perspectives below). The ultimate goal is to narrow the initially rather wide set of parameters down to the most reproducible and decisive parameters that will be incorporated into an algorithm for the prediction of donor kidney performance suitable for use in Germany, perhaps also suitable for other countries in the Eurotransplant consortium.

Table 1 Synopsis of published studies about the correlation of histopathological findings to kidney transplantation outcome
Full size table

In the following text, all grades of recommendation are given as either unconditional (strong evidence, no important drawbacks) or conditional (weaker evidence, serious potential drawbacks) following the GRADE guidelines by the Scottish Intercollegiate Guidelines Network [11].

Recommendations

Biopsy procedure

Over the last decade, most biopsies sent by the DSO have been needle biopsies containing mostly medulla and insufficient cortex (unpublished data), likely related to underestimating the feed of the biopsy gun when used on the backtable. Wedge biopsies have the disadvantage of underrepresenting arteries [12] and overrepresenting subcapsular scars. Thus, the data from studies with needle and those from studies with wedge biopsies can give conflicting results. We therefore strongly suggest to use only a single method to create a coherent database.

Punch biopsies with a 3-mm skin biopsy puncher as published from a group in Vienna, Austria, can reliably yield cortical tissue cores with sufficient arteries and glomeruli and with minimal trauma to the donor kidney [13]. Although the results from this technique have not been validated in other institutions, we recommend this technique based on personal experiences (unpublished data) and since the experience in the past with core biopsies taken under suboptimal circumstances and without microscopy to check for the quality of the biopsy core was not satisfactory (conditional recommendation). The biopsy should be performed on the backtable after organ retrieval. To prevent bleeding, the biopsy site should be marked with a single stitch with polydioxanone suture to ensure proper surgical care by the transplant surgeon after organ transplantation and reperfusion.

Biopsy techniques for the evaluation of tumours should be adapted to the site and size of the respective tumour. If an excisional biopsy is attempted, biopsies should be shipped intact to facilitate examination of resection margins (conditional recommendation).

The biopsy technique and the time of biopsy (in donor, backtable, after implantation) shall be noted.

Biopsy shipment

Donor kidney biopsies for the evaluation of parenchymal quality should be immediately placed and shipped in at least 20 ml of 4 % buffered formalin, ensuring optimal fixation of the tissue. We discourage use of other fixatives, transport in saline or on wet gauze.

In contrast, tumour biopsies should be sent on wet gauze in tight containers. Shipping on formalin or other fixatives is discouraged, since tumour biopsies can be examined as frozen sections (conditional recommendation).

For both types of biopsies, standard transport packaging with tight inner and an outer containers and absorptive material in between is to be used.

Clinical data

Clinical data about the donor, the indication for biopsy and about the macroscopic appearance of the donor kidney should be submitted in written form and online over the DSO database servers to the pathologist. The minimal dataset is detailed in Table 2. Failure to release this information to the pathologist can severely impair the histopathologic evaluation.

Table 2 Clinical data
Full size table

Biopsy processing

Parenchymal biopsies are to be embedded into paraffin with rapid techniques. In our own experience, this is also possible with 3-mm punch biopsy cores. Reservation of minute amounts of the cortex for electron microscopy is optional but not recommended in order to deliver as much tissue as possible for the initial assessment. At least two haematoxylin and eosin (H&E) and periodic acid Schiff (PAS) stained level sections should be evaluated for the immediate assessment. It is recommended to prepare additional stains like Jones silver stains or acid fuchsin orange G stains for deferred final assessments. It is left to the discretion of the pathologist to order additional stains like congo red or immunohistochemical stains for immunoglobulins or complement to reach a final definitive diagnosis (unconditional recommendation).

Tumour biopsies are to be processed according to general guidelines for the intraoperative evaluation of kidney tumours [14]. Resection margins are to be inked [15] (unconditional recommendation). According to our own data, a reliable assessment of the dignity of the tumour can be rendered in over 90 % of cases on frozen sections (unpublished data). Again, it may be necessary to use special stains or techniques to reach a final diagnosis for certain tumours.

Histopathologic evaluation

The quality of the histopathological report seems to strongly depend on the experience of the reporting pathologist with kidney biopsies [16]. However, the current shortage of dedicated renal pathologists in Germany does not allow the evaluation of all donor organ biopsies by specialists. Nevertheless, the German pathology community acknowledges the need for nephropathologists, and whenever possible, donor kidney biopsies should be evaluated by a specialist.

For the evaluation of parenchymal quality, a sufficient biopsy should contain at least three arteries and 20 glomeruli. Twenty glomeruli seems to be a good compromise between the minimal number of 25 glomeruli for accurate prediction of transplantation outcome advocated by one study [17] and the number of 10 glomeruli derived from another study [7] as well as the number of glomeruli obtainable by a typical punch biopsy [13] (conditional recommendation).

In order to ensure compatibility and comparability with future biopsy findings after transplantation and in accordance with Banff recommendations [18], all current Banff components [19–21] should be given (conditional recommendation). In addition, fibrosis of smaller preglomerular vessels (fib) should be assessed semiquantitatively as absent, minimal, moderate or severe (fib0, fib1, fib2, fib3) (conditional recommendation).

Diabetic changes are to be classified according to the Renal Pathology Society (RPS) [22]. Other vascular, glomerular or tubulointerstitial diseases shall be classified according to the relevant classification schemes endorsed by the RPS (conditional recommendation).

A synopsis of all histopathologic parameters is found in Table 3.

Table 3 Histopathological parameters
Full size table

Glomeruli

Several authors have tried to link the percentage of globally sclerosed glomeruli to transplantation outcome. More than 10–20 % of globally sclerosed glomeruli seem to be associated with unfavourable outcome [1, 7–9, 17, 23–27]. Glomerulitis as in the Banff component g [21] is only rarely found in pretransplantation biopsies and at least the leukocyte component should disappear after flushing with preservation solution. Nevertheless, it should be included in the report also to facilitate incorporation of zero-hour biopsies after reperfusion into the database.

The following findings shall be documented (conditional recommendation):

  • Total number of glomeruli

  • Number of globally (>50 % of an individual glomerulus) sclerosed glomeruli

  • Number of segmentally (≤50 of an individual glomerulus) sclerosed glomeruli. Type of segmental glomerulosclerosis (collapsing, tip lesion, perihilar, cellular or not otherwise specified) according to D’Agati et al. [28]

  • Banff components g, mm and cg [19, 21]

  • Mesangioproliferation (>3 mesangial cells per mesangial field) as focal or diffuse

  • Diabetic nephropathy according to the RPS classification (I, IIA, IIB, III or IV) [22]

Vessels

Not only arterial changes, in particular arterial intimal fibrosis [1, 7, 26, 29–31], but also arterial and arteriolar hyalinosis have been linked to an inferior transplantation outcome [17, 29, 32].

Again, the authors realise that peritubular capillaritis is an extremely rare finding in pretransplantation biopsies and could also be influenced by flushing with preservation solution. Nevertheless, it should be included in the report for the reason given above.

The following arterial/arteriolar findings shall be documented (grade of recommendation B to good practice point):

  • Number of arteries (exact number, if <11; otherwise, as >10)

  • Severity of intimal fibrosis according to Banff component cv [19]

  • Severity of hyalinosis according to Banff component ah [19]

  • Severity of fibrosis of small preglomerular vessels (fib) as absent, minimal, moderate or severe (fib0, fib1, fib2 or fib3)

  • Arteriitis according to Banff component v [19]

  • Microthrombi as glomerular and/or preglomerular, focal or diffuse

  • Peritubular capillaritis according to Banff component ptc [20]

Tubuloinsterstitium

A link between the extent of cortical interstitial fibrosis and tubular atrophy and transplantation outcome was only found in some reports [1, 8, 24, 27]. In contrast, the grading of cortical acute tubular damage has recently been examined in the USA. The authors have found a moderate association with DGF only for donors after cardia death [33]. Nevertheless, an attempt for semiquantitative assessment of cortical acute tubular damage should be made (grade of recommendation good practice point).

The following findings shall be documented (grade of recommendation C to good practice point):

  • Extent of cortical interstitial fibrosis and tubular atrophy according to the Banff components ci and ct.

  • Severity of cortical acute tubular damage should be quantified as no damage, damage but without true epithelial necrosis, damage with focal true epithelial necrosis and damage with diffuse true epithelial necrosis (atd0, atd1, atd2, atd3).

  • Interstitial nephritis as present or absent. When present with the descriptors neutrophilic or non-neutrophilic, chronic and/or active. The infiltrate shall be graded according to the Banff components i and t.

  • Nephrocalcinosis as present or absent.

Tumour evaluation

Any tumour entity, also as incidental findings in the biopsy, shall be reported following the standards of the most recent classification of the World Health Organisation (WHO). Furthermore, a statement shall be made whether a histopathologic correlate explaining a tumour was present or absent.

Communication of findings

All findings shall be communicated as soon as technically feasible online by direct entry into the database of the DSO under the name of the responsible pathologist, the local archival number, the pathologist’s affiliation and a telephone number to enable further in-depth discussion of the case. The contact telephone number of all clinicians involved will be made accessible to the pathologist online in the DSO databank to facilitate communication.

Findings of the final assessment are to be communicated electronically, by fax or by mail.

Conclusions

The authors acknowledge that there is currently no generally accepted standard for the biopsy procedure and the reporting of donor kidney biopsies and published studies give different, sometimes conflicting results. This could be due to two reasons: (1) histopathology of a donor organ does not allow prediction of the outcome or (2) the current database is insufficient, at least partially due to differences in the indication, the biopsy procedure and reporting of histopathological findings.

It cannot be overemphasised that our recommendations are only meant to standardise and to optimise the biopsy procedure, the handling, evaluation, the communication and documentation of pretransplant biopsies from deceased kidney donors. Because of insufficient data, our recommendations cannot guide the acceptance or discarding of donor kidneys. However, the standardisation and optimisation we are aiming at are a prerequisite for a meaningful collection and scientific exploitation of data derived from the evaluation of donor biopsies. Only when we will have compiled and analysed a large and coherent dataset will we be able to arrive at evidence-based, standardised procedures and will we be able to assess the clinical value of pretransplantation donor kidney biopsies in Germany.

Our recommendations will be radically different from the approach taken by the Implantation Biopsy Banff Working Group installed at the 2011 Banff meeting in Paris [34]. The Banff Working Group will advocate frozen sections from wedge biopsies, scoring only the Banff component cv (arterial intimal fibrosis), the percentage of globally sclerosed glomeruli and the extent of cortical interstitial fibrosis and tubular atrophy (H. Liapis, personal communication). The Banff Working Group’s approach will be a simplified one, and their parameters can be reported by general pathologists without special training in renal pathology. In contrast, our recommendations will enable more detailed assessment, e.g. microthrombi, in our view, almost always resulting in delayed graft function (unpublished data), will be apparent with our recommendations but is virtually impossible to detect on frozen sections. Certainly our classification requires simplification and some or most of the parameters listed above will ultimately be revealed as useless, due to high interobserver variation, due to weak associations with transplant outcome or both. Our recommendations are by no means final, and they are devised as the start of a process and will be subject to periodical evaluation, discussion and revision under the tutelage of the DGP, the DSO and the DTG.

This approach requires central collection of histopathologic data along with data about the donor, the recipient and about the outcome. Such a comprehensive collection would have to strictly adhere to the German national laws of data privacy protection. The resulting database could enable implementation of a self-learning computerised algorithm. Such an algorithm could be trained to predict e.g. the likelihood of delayed graft function for a given match of donor and recipient in real time upon entry of clinical and histopathological data. Furthermore, such an algorithm could be trained to decide, upon entry of clinical data, if a histopathologic examination would be necessary at all for more precise outcome prediction. The likely overrepresentation of marginal donor kidneys in this database could be balanced by the addition of clinical datasets from unselected donors, recipients, histopathology and transplant outcome derived from already existing collections of zero-hour biopsies.

The resulting database could be of immense use for clinical decision making, patient-oriented research and transparency in German transplantation medicine.