Role of Donor Kidney Biopsies in Renal Transplantation

Recentyears have seen an increasing use of marginal donors to expand the organ poolavailable for renal transplantation (1–5).Donors considered in the marginal category include those with age >55years, hypertension, diabetes mellitus, acute tubular necrosis, disseminatedintravascular coagulation, prolonged cold ischemia time, and non-heart-beatingdonors. Patient and graft outcome obtained with such suboptimal donors hasbeen comparable to that obtained with ideal donors in some studies (6, 7),but significantly worse in others (2, 8–10).These variable results presumably reflect the use of organs with differentdegrees of functional reserve. It has been estimated that donor factors canaccount for 35–64% of the variation in recipient serum creatinine andcreatinine clearance after transplantation (11, 12).Hence, it stands to reason that demonstration of satisfactory donor kidneyfunction before accepting an organ for transplantation would improve bothshort- and long-term graft function. Donor assessmentshould begin with a review of clinical data, but in cases of traumatic deathadequate prior medical records are not always available. Some centers haveused an arbitrary age cut off to exclude donors with senilearterionephrosclerosis. However, this is not entirely a satisfactory approach,because of individual variability in the rate at which kidney tissue ages.Thus, the percentage of sclerotic glomeruli in human kidneys varies between0.2–16.7% at age 55 years and 1.5–23.0% at age 75 years (13).Data from a study conducted at the University of Pittsburgh indicated that17/30 (57%) donors aged 60–75 years had 0–10% glomeruli sclerotic:clearly a decision to reject these donors based on age alone would have beeninappropriate (3).Conversely, mild histologic abnormalities can be present in youngerindividuals much more commonly than is generally appreciated. Arteriolarhyalinosis has been reported in 25- to 34-year-old subjects, and considered tobe a marker for early onset atherosclerotic disease (14). Laboratoryevaluation of donor renal function is important, and should include urineexamination as well as blood chemistry. It should be kept in mind that mildproteinuria can occur secondary to glomerular or tubular ischemia reflectingagonal changes occurring before death. Acceptable cut off values forproteinuria used by different centers range between 0.5–3.0 g/24 hr (15, 16).Blood urea and serum creatinine are readily available parameters for theassessment of renal function, but can rise significantly due to conditionssuch as renal hypoperfusion and acute tubular necrosis, which do not per secontraindicate transplantation. Creatinine clearance has also been used forscreening of donors, and is superior to serum creatinine in that it is notaffected by donor age, muscle mass, or obesity, However, clear-cut guidelineson the use of creatinine clearance as a criterion for donor selection have notyet been developed. Some authors have suggested a donor creatinine clearancemeasurement >60–70 ml/min for accepting marginal organs for singlekidney transplantation (15, 17).In contrast, others investigators have recommended double kidneytransplantation when the donor creatinine clearance is less than 90–100ml/min (18).Allograft function cannot be simply predicted by evaluating the donorcreatinine clearance, because of multiple posttransplant variables such asacute tubular necrosis, antibody or cell-mediated rejection, and calcineurininhibitor nephrotoxicity. Another confounding factor is the occurrence ofcompensatory renal parenchymal hypertrophy, when the donor nephron mass isinsufficient to meet the metabolic needs of the recipient. Compensatorychanges have been shown to result in an approximately 20% rise in estimatedcreatinine clearance in the allograft kidney within 4–6 months oftransplantation (16). Theremainder of this review will focus on the role that a pretransplant biopsycan play in helping to define the structural integrity and functional reserveof a donor kidney under consideration of transplantation. A biopsy should beconsidered mandatory when the donor in question is in the marginal category.At Pittsburgh, we have set an arbitrary cut-off age of 55 years, beyond whichall donors are biopsied to evaluate the severity of senilearterionephrosclerosis. A strong case can be made to include apretransplantation or postperfusion biopsy in the routine work up of alldonors, irrespective of age and clinical setting. This would provide baselineanatomic data with which future biopsies can be compared. Preexisting lesionssuch as capillary thrombosis, arteriolosclerosis, glomerulosclerosis, andinterstitial fibrosis can be recorded, so that the occurrence of the samelesions in posttransplantation biopsies is not misconstrued as evidence ofcalcineurin inhibitor nephrotoxicity or chronic allograft nephropathy. Lack ofknowledge about the extent of preexisting changes in a donor kidneycomplicates the interpretation of posttransplantbiopsies. Biopsy techniques vary from institution toinstitution. I prefer a generous wedge biopsy about 1-cm long and 0.5-cm deep.This suggested size ensures that at least half the cortical depth is availablefor evaluation, and minimizes erroneous conclusions due to superficialsubcapsular scarring secondary to senile arteriosclerosis. Some centers preferthat both a wedge and a needle biopsy be performed to provide assurance thatthe deep cortex has been adequately sampled. A needle biopsy alone may notpermit reliable assessment of the extent of glomerulosclerosis due to limitedsampling. One study has suggested that sample adequacy be defined by thepresence of a minimum of 25 glomeruli (19).This contention was based on the observation that a statistically significantrelationship between percent glomerulosclerosis and graft loss was observedonly if biopsies with more than 25 glomeruli were analyzed. In another study,the relationship between percent glomerulosclerosis and graft function wasfound to hold irrespective of the number of glomeruli present at biopsy (2).However, in that study this relationship was lost on multivariate analysis ifa correction was made for donor age. This led the authors to state that if thedonor age is known, data on glomerulosclerosis do not add any additional vitalinformation. However, as pointed out earlier, age-associated changes in thehuman kidney are extremely variable and can not be predicted without a biopsy.Additionally, a biopsy can detect the presence of previously undocumentedchronic diseases such as hypertensive or diabetic nephropathy, and chronictubulointerstitial nephritis. Urgent histologicalprocessing of donor biopsies is needed, when the decision to use the donorkidney is contingent on the morphologic findings. Because prolonged coldischemia can adversely affect long-term graft function, the biopsies need tobe interpreted as soon as possible. Rapid processing protocols can allowpermanent sections to be available for reading within 2 hr. Consistentlyproviding this level of service, however, necessitates that both ahistotechnologist and an anatomic pathologist to be on call round the clock.As an alternative, a frozen section service with only a pathologist being oncontinuous call can be offered. Frozen section morphology is adequate torecognize sclerotic glomeruli, advanced interstitial fibrosis, andarteriosclerosis. However, freezing artifacts can lead to interstitialwidening, which can be confused with fibrosis, if one does not insist ondemonstrating a definite collagenous matrix. Retraction of tubular epitheliumfrom the basement membranes makes it difficult to recognize tubular atrophy.Frozen sections are also not reliable for assessment of mesangial cellularity,glomerular capillary wall thickening, and diabetic lesions such as smallcapsular drop lesions or early Kimmelstein-Wilson nodules. Gross thrombosiscan be recognized at frozen section, but small fibrin thrombi in thecapillaries are more difficult toevaluate. Interpretation of a kidney biopsy from a donorwith senile arterionephrosclerosis or other chronic renal disease calls for asemiquantitative evaluation of the degree of glomerulosclerosis,arteriosclerosis, and interstitial fibrosis present. The use of Banff criteriafor grading chronic allograft nephropathy is suggested to ensure center tocenter uniformity in this assessment (20).If most of the glomeruli are patent, and there is only mild arteriosclerosisand interstitial fibrosis present, the donor kidney is suitable for use.However, the extent of acceptable chronic changes within the donor kidney hasnot yet been rigorously defined. A widely accepted empiric rule is thatkidneys with more than 20% sclerotic glomeruli not be used (8).At Pittsburgh, surgeons are also hesitant to use any kidney with more thanmild interstitial fibrosis (more than 25% of cortical area affected) or mildarteriosclerosis (more than 25% luminal occlusion). Glomerular, interstitial,and vascular lesions in any given biopsy are frequently proportional to eachother, even though this interrelationship is somewhat imperfect (21).Hence, we have taken the approach that moderate or severe changes in any ofthe major anatomic compartments in a donor kidney should contraindicatetransplantation. Recently, it has been shown that the maximal planar area ofthe nonsclerotic glomeruli is also a predictor of long-term graft function (22). Severalinvestigators have studied interobserver variability in grading morphologicchanges in donor biopsies. Pokorna et al. reported moderate to goodreproducibility with calculated weighted kappa scores of 0.66 for percentglomerulosclerosis, 0.78 for interstitial fibrosis, and 0.83 for arteriolarhyalinosis (2).Wang et al. addressed this issue by (1) comparing histological changes inpaired baseline biopsies from the same donor, and (2) comparing baseline donorbiopsies with sequential posttransplant biopsies from the same recipient (19).Using linear regression analysis, the precision of estimating percentglomerulosclerosis in paired biopsies was good only if analysis was restrictedto biopsies with more than 14 glomeruli (r=0.83 for paired biopsies andr=0.56 for sequential biopsies). The κ statistic for arteriolarhyalinosis was 0.55 for paired biopsies and 0.38 for sequential biopsies.Discrepancies in grading arteriolar hyalinosis were found in 10% of pairedbiopsies and 20–30% of sequential biopsies. Sund et al. reported poorreproducibility in the grading of arteriosclerosis and arteriolar hyalinosisin sequential biopsies, based on calculated kappa scores of 0.046 and 0.122,respectively (22).These disappointing results presumably reflect variation in the distributionand severity of vascular lesions in this patient population. The lesions weremore pronounced in the pretransplant biopsy compared to the posttransplantbiopsy. It was suggested that this difference resulted from a propensity ofvascular lesions to affect deeper vessels, which are more likely to be sampledwhen a biopsy gun is pointed directly at the surface of a donor kidney. Insupport of their contention, the authors pointed out that thecortico-medullary junction was indeed more often sampled in biopsies obtainedbefore transplantation. Several studies have validatedthe clinical utility of donor biopsies by formal statistical analysis. Seronet al. examined postperfusion biopsies, and showed a correlation betweeninterstitial fibrosis and serum creatinine measured 12 months posttransplant (9).Leunissen et al. showed a correlation between a histological chronicity scoreobtained at postperfusion biopsy and creatinine clearance measurementperformed 3 months later (23).Lehtonen et al. found the chronic allograft damage index in a donor biopsy topredict long-term graft function (24).Gaber et al. showed that postperfusion biopsies with >20%glomerulosclerosis (n=8) were associated with an 88% incidence ofdelayed graft function (7/8 grafts), 38% graft loss (3/8 grafts), and a meanserum creatinine of 2.6±0.1 mg/dl at 6 months (8).Based on this data, it was suggested that kidneys with >20%glomerulosclerosis not be used for transplantation. However, this conclusionwas derived from a study group of only eight patients with an unusually highpercent glomerulosclerosis (mean 39±6%). The control group of patientsused for comparison had significantly lower glomerulosclerosis (8±1%).Pokorna et al. described a 3-year graft survival of 74.7% in 67 patients with20.0–47.6% glomerulosclerosis, but 11% of these recipients had primarynon-graft function, and a mean 1-year glomerular filtration rate of 41.4ml/min (2).Several studies have demonstrated a relationship between donorarteriosclerosis and posttransplant function (25–27).Hyaline changes in the smaller arteriolar sized vessels also correlate with 1year serum creatinine (28) and rate of graft failure (19). Thereason why interstitial fibrosis, glomerulosclerosis, arteriolar hyalinosis,or arteriosclerosis have variably been identified as the critical parameter indifferent studies is probably the result of patient selection andmethodological considerations. For example, in one study where donorglomerulosclerosis, but not interstitial fibrosis, was found to predict graftfunction, cases with >55% and <55% interstitial fibrosis were comparedwith regards to the incidence of satisfactory graft function defined simply asa patient being alive without maintenance dialysis (19).In a second study, interstitial fibrosis was found to be predictive, ifbiopsies showing no interstitial fibrosis were compared with those showing anylevel of interstitial fibrosis, and graft function was assessed by calculatedcreatinine clearance (3).Failure to detect the effect of interstitial fibrosis in some studies may alsopartly reflect the patchy nature of this lesion, which in turn, may be due tothe patchy nature of arteriosclerosis and arteriolar hyalinosis in thekidney. In contrast to the literature discussed above,one can also find studies that fail to find any correlation between donorbiopsy findings and posttransplantation graft function (7, 22, 29–32).This is surprising given the intuitively expected relationship betweenanatomic architecture and physiological function in the kidney. Closeranalysis of many of these studies reveals methodological problems such as (1)small numbers of patients, (2) insufficient histological detail for criticalevaluation, (3) studies limited to biopsies with only mild histologicalchanges, (4) lack of correction for variables such as prolonged cold ischemiaor acute rejection, and (5) use of only crude patient or graft survival ratesin evaluating outcome (1, 7).In some clinical settings, the expected effect of donor histology can probablybe overshadowed by other confounding clinical variables. Thus, many surgeonsprefer to give kidneys from older donors to older recipients, who have aweaker immune system. This may result in lower rejection and reasonable graftsurvival, despite changes of senile arterionephrosclerosis in the donor organ.One study has suggested that improvements in medical care have now reduced theimportance of donor age as a critical factor in renal transplantation (33). Thepreceding discussion has focused primarily on donor biopsies performed for oldage, hypertension, or donor diabetes mellitus. Clinical concern aboutpretransplant ischemic injury is another relatively common reason forrequesting a donor biopsy. Predisposing factors for such injury include ahistory of donor hypotension, use of pressors during donor medical management,prolonged cold/warm ischemia time, a non-heart-beating donor, and chronicparenchymal or vascular disease in the donor. Acute tubular necrosis, thehistological counterpart of ischemic injury, is difficult to evaluate onfrozen section, except in cases with frank coagulative necrosis or infarction.Even with ideal permanent section morphology, correlations between clinicalrenal dysfunction and histological acute tubular necrosis are imperfect. Solezet al. could not demonstrate any correlation between histological severity ofacute tubular necrosis and duration of oliguric acute renal failure in thenative (nontransplanted) kidney (34, 35).Lehtonen et al. found that chronic changes in the donor biopsy did notcorrelate with immediate posttransplant graft function (24).A similar lack of correlation has been observed with donor vascular disease (27).However, others have reported that histological scoring for acute tubularnecrosis (2, 26, 36) or apoptosis (37) predicts delayed graft function. There is evidence that prolonged coldischemia and delayed graft function predisposes to vascular rejection inkidneys derived from older donors (35).Whether delayed graft function adversely affects long-term graft survivalindependently of rejection is controversial (36). Donorbiopsies performed in the setting of disseminated intravascular coagulationneed to be evaluated for the extent of microvascular injury. Organs withdiffuse and extensive glomerular thrombosis should be discarded. However, thepresence of scattered capillary thrombi present in a minority of glomerulidoes not necessarily contraindicate transplantation. When the donor serumcreatinine is normal or marginally elevated, successful transplantation hasbeen reported. Isolated fibrin thrombi can apparently be dissolved by anintact fibrinolytic system (31, 37),although this may result in a transient microangiopathic hemolytic anemia in afew instances (38).Mate kidneys recipients from the same coagulopathic donor can have differentgraft outcomes due to variations in pre- and posttransplant factors (39). Occasionally,pretransplant or postperfusion biopsies show changes consistent withglomerulonephritis, and allow the glomerular disease in the allograft kidneyto be traced back to the organ donor. The risk of this scenario is probablythe highest for IgA nephropathy, a disease with high prevalence in somegeographic regions. Based on isolated case reports in the literature, it wouldappear that mild glomerular changes in a donor biopsy can probably be ignored.Thus, it has been documented that modest donor-derived IgA deposits do notcause significant graft dysfunction, and can spontaneously resolve with time (40, 41).Similar observations have been made regarding donor-derived postinfectiousglomerulonephritis, membranoproliferative glomerulonephritis type I and lupusnephritis (42–44).Focal segmental sclerosis attributable to donor disease has been shown not toprogress in the posttransplantation period (31). Thefinal indication for a donor kidney biopsy is the presence of a grosslyvisible nodule noticed during harvesting of the organ. When histologicalexamination shows a benign cyst, leiomyoma or angiomyolipoma, it is safe toproceed with transplantation. However, finding a small epithelial neoplasm cangenerate dilemmas that may be difficult to resolve, particularly when a highgrade carcinoma is not demonstrated. The distinction between a so-called renaladenoma and a small low grade renal cell carcinoma is arbitrary, andtraditionally based on the size of the lesion, although it is now increasinglyrecognized that lesions of any size can metastasize. If the donor lesion issmall (less than 0.5 cm) and completely excised, the risk of residual orrecurrent carcinoma in the recipient is probably extremely small. Dr. IsraelPenn has reported six cases, where wide excision of the donor nodule led to anuneventful course documented by up to 186 months of posttransplantation followup (45-48).The rare occurrence of posttransplant renal allograft carcinoma, despite theestimated 7–25% incidence (based on routine autopsy data) of small in donor also that the use of such be at least in the of recipient this is a controversial and some centers organs with small epithelial a kidney biopsy is in the clinical of who are being evaluated for renal transplantation. In it the of care to a baseline biopsy from kidneys irrespective of the medical can arterionephrosclerosis and renal disease in the donor. Lack of about of preexisting changes in a donor kidney the allograft nephropathy and hyalinosis in posttransplantbiopsies. donor biopsies, an should be made to severity of glomerulosclerosis, interstitial fibrosis, arteriolar hyalinosis present. of available evidence organs with glomerulosclerosis and mild interstitial fibrosis give acceptable If the biopsy changes the of a suboptimal organ with has to be the of the patient by kidney transplantation can also beconsidered in the (16).