Renal Histopathological Lesions After Orthotopic Liver Transplantation (OLT)

American Journal of Transplantation 2005; 5: 1120–1129 Copyright C Blackwell Munksgaard 2005 Blackwell Munksgaard doi: 10.1111/j.1600-6143.2005.00852.x

∗ Evangeline Pillebouta, , Dominique Nochyb, Introduction Gary Hillb, Filomena Contic, Corinne Antoinea, Yvon Calmusa and Denis Glotza With the increased survival of patients with liver transplants, the appearance of chronic renal failure (CRF), some- aHopital Saint Louis, Nephrology and Kidney times requiring dialysis or renal transplantation, has be- Transplantation, Paris, France come a frequent complication and an important cause of bHopital Europeen Georges Pompidou, Histopathology, post-operative morbidity and mortality (1–3). The cumu- Paris, France lative incidence of CRF has been recently estimated at cHopital Cochin, Surgery and Liver Transplantation, 18.1% at 5 years (4). Paris, France ∗ Corresponding author: Evangeline Pillebout, Aside from the hepatorenal syndrome in which functional [email protected] renal failure is treated by orthotopic liver transplantation (OLT) (5,6), in all other situations of hepatic insufficiency, Liver transplant recipients are at risk of chronic renal factors responsible for the appearance or worsening of failure (CRF), customarily considered to be secondary CRF following transplants are still poorly understood. The to CsA/FK506 nephrotoxicity. We have examined renal appearance of CRF following transplantation suggests fac- biopsies from 26 liver transplant recipients with CRF. tors related to the transplant, particularly immunosup- Before OLT, 5 patients had CRF, 8 were diabetic and 9 hypertensive. Renal biopsies were performed at a pressive treatment. Quite different is the worsening of a mean of 5 years after liver transplantation. Mean SCr pre-existing CRF. This pre-existing disease is often under- was then 212 lmol/L, proteinuria was 1 g/24 h. Twelve estimated because of the significant muscle loss these patients were diabetic and 25 hypertensive. Histology patients typically manifest. Serum creatinines (SCrs) are revealed impressive renal destruction, with a mean of seemingly normal and the renal insufficiency is difficult 45% interstitial fibrosis and 45% glomerular sclerosis. to demonstrate (7). In order to avoid these complications All biopsies showed severe arteriosclerosis. CRF can be some authors have proposed performing a double hepatic- attributed to four associated primary lesions: (i) spe- renal transplantation (8). cific chronic CsA/FK506 arteriolopathy; (ii) typical diabetic nephropathy; (iii) acute or chronic thrombotic microangiopathy attributed to CsA/FK506 or a -IFN and The majority of authors (1,4,9–11), have attributed the (iv) tubular changes related to administration of hy- development of CRF post-transplantation, to calcineurin droxyethylstarch. At the end of the follow-up, after a inhibitor therapy, with either cyclosporine or tacrolimus. mean of 6.4 years, 12 patients required dialysis, 13 had The nephrotoxicity of calcineurin inhibitor has been clearly CRF and only 1 had normal renal function. Thus, CRF in demonstrated in cardiac (12,13) and renal (14–16) trans- OLT recipients is more complex than originally thought plants in which the therapeutic doses are higher than in and should not be classified as anti-calcineurin nephro- hepatic transplants. toxicity without further investigations, including renal histology. These investigations have therapeutic po- tential, that is, they may lead to a more aggressive In addition, in patients transplanted for hepatitis-C-related treatment of hypertension and/or diabetes. cirrhosis, intense viral replication post-transplant, with the risk of destruction of the new liver, sometimes requires Key words: Interferon (IFN), liver transplantation, ne- treatment with alpha interferon (a-IFN), itself nephrotoxic oral FK506 nephrotoxicity, post-transplant diabetes, (17–21). Finally, the glucose dysregulation in chronic hep- renal biopsy, renal failure atic insufficiency, leading in some to insulin-dependent ‘hepatogenous diabetes’, recognized as early as 1898 in Received 2 September 2004, revised 14 December 2004 cirrhotic patients (22) can also play a role in the renal and accepted for publication 5 January 2005 dysfunction.

This new study, extending a previous one (23) to 26 patients with hepatic transplants and CRF demonstrates that renal involvement is often severe and that renal destruction

1120 Orthotopic Liver Transplantation is in fact multifactorial. There are, to be sure, not only the tions), the presence of diabetes (defined as need for oral hypoglycemic specific lesions of immunosuppressive treatment, but also agents or insulin) and biologic evidence of hemolysis (haptoglobin, LDH lesions related to the volume-expansion products used in and schizocytes). The last 10 patients to be studied had measurement of France in patients awaiting transplant and to a-IFN, as well glomerular filtration rate by EDTA clearance. as lesions of diabetes and hypertension. Renal biopsy processing This study will renew the debate of possible double liver- Standard percutaneous biopsies were performed in 25 patients and surgical renal transplantions. It will also permit fairly precise thera- biopsy in 1. Specimens were fixed in alcoholic Bouin’s fixative with standard peutic suggestions to slow the often rapid evolution toward embedding and stained. Routine immunofluorescence studies were also end-stage renal disease (ESRD), based on the discovery of done using polyclonal antibodies to IgA, IgG, IgM, fibrin, albumin, C3, C4, certain renal lesions whose significance had not been eval- C1q (Behringwerke, Marburg, Germany) and anti-j and lambda light chain uated up to now. antibodies (DakopattsA/S, Glostrup, Denmark) by standard methods.

All of the slides were reviewed by two experienced pathologists (DN, GH), Patients and Methods blind to the clinical and laboratory data. All biopsies contained at least nine glomeruli.

Patients This study included 29 patients who received OLT and then underwent Histologic lesions renal biopsy between 1999 and 2003. All patients were followed in the liver A variety of basic lesions were evaluated. Sclerotic glomeruli and glomeruli transplantation unit of the Hopital ˆ Cochin and in the renal transplantation unit with evidence of focal segmental glomerulosclerosis (FSGS) were counted of the Hopital ˆ Europeen ´ Georges Pompidou, both in Paris. Three patients and expressed as percentages, and the degree of interstitial fibrosis was showing renal lesions related to their hepatic disease (hepatitis-C-related estimated as a percentage of the parenchyma. Other lesions were eval- cryoglobulinemia) were excluded. uated on a semi-quantitative scale of 0–3+.These included arterial, arteriolar, renal tubular and other lesions. Diabetic glomerulosclerosis was Twenty one were men (82%), 5 were women (18%). All but one (Asian) sought by light microscopy and also on IF by specific linear staining for IgG were Caucasian, including 4 (15%) from North Africa. The mean age was and albumin along glomerular capillary walls and tubular basement mem- 57 ± 2 years (range: 43–70 years). The causes of the liver disease were as branes. Calcineurin inhibitor toxicity was sought in the form of specific ar- follows: hepatitis C (n = 16), hepatitis B (n = 5, of whom 2 were coinfected teriolopathy with medial smooth muscle necrosis (24). Because they can- with both hepatitis B and C), alcoholic liver disease (n = 9 associated with not be distinguished from one other, renal lesions due to cyclosporine A hepatitis C or B in 7). The mean waiting period on the transplant list was or tacrolimus were grouped (25). Striped fibrosis was excluded from con- 10 ± 4 months (range: 0–49 months). During surgery, the patients re- sideration in this setting because of the multiplicity of other causes of in- ceived a mean of 1892 ± 238 mL (range: 0–4500 mL) of colloids, interstitial fibrosis. Thrombotic microangiopathy (TMA) was also evaluated cluding hydroxyethylstarch, and 6462 ± 651 mL (range: 0–11 500 mL) of separately, because it may have causes other than calcineurin inhibitor tox- crystalloids. icity. Also sought in all biopsies were interstitial fibrosis and tubular atrophy and calcifications, tubular and interstitial foam cells associated with over- The immunosuppressive drugs used immediately post-transplant were load by hydroxyethylstarch (Elhoes®)used for volume expansion. The lat- prednisone for all patients, associated with cyclosporine A (Neoral®, No- ter were distinguishable from tubular lesions of calcineurin inhibitor toxicity vartis Pharma SAS, Rueil-Malmaison, France) for 18 patients or tacrolimus by their coarse vacuolization and pyknotic nuclei, rather than the fine ‘os- (FK-506, Prograf®,Fujisawa SARL, La-Celle-Saint-Cloud, France) for 7 oth- motic’ granularity of calcineurin toxicity, and their occurrence primarily in ers. In addition, 7 patients received azathioprine (Imurel® GlaxoSmithKline, atrophic tubules and particularly in the interstitium, rather than in function- Marly-le-Roi, France), and 5 patients received mycophenolate mofetil (Cell- ing nephrons (Figure 2B). Arteries were evaluated according to the pres- cept® Roche, Neuilly-sur-Seine, France). ence and extent of arteriosclerosis in the form of intimal fibroplasia and thickening often with medial atrophy, and the presence of smooth muscle hypertrophy. Arterioles were evaluated for the presence of medial necro- Liver graft survival sis (as a sign of calcineurin inhibitor toxicity), and of intimal subendothelial Seven hepatitis C patients had a recurrence (50%) with histologic signs hyaline deposits (classical hyaline arteriolosclerosis of hypertension and dia- of viral replication necessitating treatment with ribavirine and a-IFN. Two betic nephropathy). In the great majority of instances these lesions could be patients had a second hepatic transplant, 1 patient at 1 year 8 months after readily distinguished from one another, and it was often found that the two the initial transplant because of a chronic cholangitis, and the second patient coexisted. at 3 years 5 months after the first transplant because of an arterial thrombus of the graft. In a separate study, one author (GH) counted all of the glomeruli and arterioles in each biopsy. Glomeruli were categorized as: normal, FSGS, diabetic Clinical monitoring and follow-up glomerulosclerosis, ischemic and sclerotic. Arterioles were categorized as: Data were gathered before and after OLT, at the time of renal biopsy (RB) normal, calcineurin inhibitor arteriolopathy, hyaline arteriolosclerosis and a and at the end of follow-up. The following data were recorded: demographic combination of both types of lesions. Results were expressed in percent- information, clinical and laboratory data relative to hepatic function (alkaline ages of total glomeruli or arterioles in the biopsy. phosphatase, aspartate aminotransferase, alanine aminotransferase and total serum bilirubin) and renal parameters (SCr, 24-h protein excretion and hematuria). Renal failure was defined as a creatinine clearance (CrCl) of Statistical analyses <60 mL/min, as calculated by the Cockcroft formula. Equally documented Demographic, clinical, biologic and histologic variables were analyzed using were the presence of hypertension (systolic pressure >150 mmHg and/or Statview 5.0 software (SAS Institute Inc. Cary, NC), with results expressed diastolic pressure >90 mmHg or the need for anti-hypertensive medica- as mean ± standard deviation. Categorical variables were compared by

American Journal of Transplantation 2005; 5: 1120–1129 1121 Pillebout et al. chi square tests. Others variables were compared by non-parametric tests nephrotic syndrome. Two patients diabetic before trans- (Mann-Whitney and Kruskal-Walis), with values of p ≤ 0.05 regarded as plantation could be weaned from insulin in the weeks significant. post-transplant. By contrast, 6 other patients (23%) developed insulin-dependent diabetes following hepatic trans- Results plant. Overall, at the time of RB, 12 patients (47%) were diabetic and all but 1 were hypertensive. The doses of immunosuppressive drugs were: 2.4 ± 1.2 mg/kg/day (0.7– Prior to hepatic transplantation 4.7 mg/kg/day), trough levels in the range of 122 ± 18 lg/L The renal clearance was 77 ± 5 mL/min (30–130 mL/min) (60–232 lg/L) for the 12 patients treated by cyclosporine with a mean SCr of 90 ± 4 lmol/L (65–140 lmol/L) [9.1 ± and 0.4 ± 0.06 mg/kg/day (0.02–0.08 mg/kg/day), trough 0.5 mg/L (7.4–15.9 mg/L), 1 mg/L = 8.8 lmol/L]. Five pa- levels in the range of 8 ± 1.5 lg/L (3.4–15 lg/L) for the 13 tients had overt CRF before renal transplantation. There patients treated by tacrolimus. In 6 patients cyclosporine was pre-existing diabetes in 8 patients (31%), with insulin was stopped and replaced by tacrolimus, in 5 of them be- required in 5. Nine patients (35%) were hypertensive. cause of rejection and because of nephrotoxicity in the sixth (the switch being made several months before RB). Peri-hepatic-transplantation operative period (from Day 0 to the end of the first hospitalization, mean 43 ± 3 days (16–71 days)) At the end of the follow-up ± All patients but 2 developed acute renal failure (ARF) The patients were followed a mean of 4.8 0.7 years (from in the post-hepatic transplantation. The mean CrCl was 6 months to 11.6 years) after the RB. Twelve patients (46%) 44 ± 4 mL/min (5–120 mL/min) with a mean SCr of 186 ± have gone on to ESRD requiring dialysis (Figure 1). Thir- 27 lmol/L (63–711 lmol/L). Two patients required short- teen patients (50%) had renal failure, with a mean CrCl of ± ± term dialysis. Two patients developed delayed ARF 10 45 3 mL/min (28–60 mL/min) and a mean SCr of 185 and 15 days post-surgery. Renal function improved in all 23 lmol/L (120–445 lmol/L). A single patient had normal patients in the weeks post-operatively. At discharge, the renal function with neither hematuria nor proteinuria. It mean CrCl was 55 ± 4 mL/min (16–93 mL/min) with a should be pointed out that renal function is overestimated mean SCr of 142 ± 15 lmol/L (75–380 lmol/L). However by the Cockcroft formula in these patients. The last 10 pa- CRF persisted in 14 patients, severe in 5 of them with CrCls tients had EDTA clearances. In these patients clearance cal- ± of less than 30 mL/min. culated by Cockcroft formula was 39.2 4.7 mL/min (15.1– 64.0 mL/min) but EDTA clearance was 27.4 ± 4.6 mL/min At the time of renal biopsy (9.2–42.8 mL/min). In fact, in these patients, the Cockcroft ± − Renal biopsies were performed at a mean of 4.8 ± formula overestimated function by 32 7% ( 15–55%). 0.7 years after hepatic transplant (0.5–11.6 years). The in- At the last follow-up, all patients were hypertensive and dication for biopsy was CRF in 25 of the patients, and the 12 diabetics were all insulin-dependent, though none isolated proteinuria of 1.2 g/24 h in the last patient. Re- has presented macro- or microvascular complications. nal function had declined in all 26 patients between the time of transplantation and RB. At the time of biopsy, the Renal histopathologic lesions mean CrCl was 36 ± 3 mL/min (from 5 mL/min [arbitrary The histologic lesions were extensive (Figure 2). All biop- value in dialyzed patients] to 69 mL/min) and a mean SCr sies considered, interstitial fibrosis was estimated at 45 ± of 212 ± 16 lmol/L (120–445 lmol/L). Two patients had 3% (10–70%), and the percentage of sclerotic glomeruli microscopic hematuria. Proteinuria was generally moder- was similar, 45 ± 4% (0–82%), associated with marked ate, 1.0 ± 0.3 g/24 h (0–4.5 g/24 h). Two patients had a vascular lesions. However, even at this point it was

350 12 patients

300 mol/L)

µ 1 patient 250 2 patients 200 0 patient

150

100 Figure 1: Immediate post-oper- 50 ative acute renal failure and long- Serum creatinine ( term chronic renal failure in liver 0 transplant patients. Mean delay of Post-operative End of the first Renal biopsy End of the Before OLT creatinine dosage is noted for each (14 days ± 2) Hospitalisation (4.8years ± 0.7) follow-up ± ± point. Number of dialyzed patients is (43 days 3) (6.4 years 0.7) noted upon the bars.

1122 American Journal of Transplantation 2005; 5: 1120–1129 Orthotopic Liver Transplantation

Figure 2: Renal histopathologic lesions. (A) Arterial lesions. The majority of biopsies reveal interlobular arteries showing severe, nearly obstructive arteriosclerosis associated with ischemic glomeruli and interstitial fibrosis. Masson trichrome, ×200. (B) Chronic hydroxyethylstarch nephotoxicity and focal cortical atrophy (FCA). Biopsy from a patient showing an area of focal cortical atrophy in the superficial cortex. This is associated with lesions related to the infusion of Elhoes®, with an infiltration of the tubules and interstitium by cells with micovacuolated cytoplasm and shrunken nuclei, and tubules with osmotic-nephrosis-like lesions. Masson trichrome, ×200. (C) Thrombotic microangiopathy (TMA) from a patient with arteriolar FK 506 vascular nephrotoxicity. The glomerular hilar region shows TMA with fibrous occlusion (∗). Masson trichrome, ×400. (D) Diabetic nephropathy. Nodular glomerulosclerosis from a patient with typical diabetic nephropathy. This glomerulus also shows an area of mesangiolysis in the upper lobule, surrounded by the beginning of an area of focal segmental glomerulosclerosis with vacuolated podocytes. Masson trichrome, ×200. possible to recognize specific histologic lesions suggesting the use of hydroxyethylstarch, consisting of ‘osmotic- the intervention of multiple factors in the renal destruction. nephrosis-like lesions’ of the distal and proximal tubules (26) (Figure 2B). We were surprised to find these lesions The percentage of patients having each of the major le- in 16 of our patients (61%), persisting up to 10 years af- sions is: vascular lesions 65%, tubulointerstitial lesions re- ter the intravenous use of this product. These cells, with lated to hydroxyethylstarch 61%, TMA 50%, CsA/FK506 microvacuolated cytoplasm and shrunken nuclei are prob- nephrotoxicity 46%, diabetic lesions 34% and lesions of ably laden with this product, although this has yet to be FSGS 34%. Two biopsies also showed mesangial IgA de- proven. posits, but without light microscopic evidence of glomeru- lonephritis. As shown in Table 1, all of these lesions were in- Thrombotic microangiopathy (TMA): TMA was pre- tertwined, with a given biopsy showing up to five of them. sent in 13 patients (50%), with thrombotic and organizing thrombotic lesions of the pre-glomerular arterioles in 8 pa- Arterial lesions: These were present in all of the biop- tients and fibrotic and occlusive lesions in 5 (Figure 2C). The sies, but particularly diffuse and severe in 17 patients (65%) majority of cases of TMA also had lesions of CsA/FK506 (Figure 2A). Particularly impressive was the marked hyper- nephrotoxicity, and 4 of 7 patients having taken a-IFN trophy and hyperplasia of the myocytes in inter-lobular and showed TMA, in two of them they were diffuse and fibri- arcuate arteries. Small arteries and arterioles were also nous. ‘Focal cortical atrophy’ (FCA), as described in primary markedly altered, with intimal lesions related to hyperten- anti-phospholipid syndrome (27) was present in 8 patients. sion and diabetes, to which were added the lesions of It is characterized by involvement of the superficial cor- CsA/FK506 toxicity in the media. tex with interstitial fibrosis, tubular thyroidization, sclerotic and cystic with collapsing glomeruli and occluded arterioles Chronic hydroxyethylstarch nephrotoxicity: These his- (27). This FCA may be a feature of nephroangiosclerosis or tologic lesions were first reported a decade ago, after of TMA, or both. It can be distinguished from the striped

American Journal of Transplantation 2005; 5: 1120–1129 1123 Pillebout et al.

Table 1: Renal biopsy ﬁndings in 26 hepatic transplant patients∗ Severe arterial Hydroxyethylstarch CsA/FK506 Diabetic Patient lesions nephrotoxicity TMA arteriolopathy lesions FSGS 1 ++ ++ + 2 ++ ++ + 3 ++ ++ + 4 ++ + ++ 5 ++ +++ 6 ++ ++ 7 ++++ 8 ++++ 9 ++ + 10 +++ 11 +++ 12 +++ 13 + ++ 14 ++ + 15 ++ + 16 +++ 17 ++ 18 ++ 19 ++ 20 + + 21 + + 22 ++ 23 + 24 + 25 + 26 Total 65% 61% 50% 46% 34% 34% ∗Most biopsies showed numerous intertwined lesions. Severe arterial lesions are deﬁned as diffuse and very obstructive arteriosclerosis with either medial atrophy or hypertrophy of the media.

fibrosis of calcineurin inhibitor toxicity by its uniquely sub- Clinicoanatomic correlations capsular location, rather than the medullary ray distribu- In terms of renal function, a number of correlations that tion of the latter. These 8 patients with FCA showed far might be expected were indeed confirmed. For the group worse morphologic features than those in which FCA was as a whole there were significant correlations between SCr absent. and: percentage of sclerotic glomeruli (r = 0.4339, p = 0.03), FSGS (r = 0.5301, p = 0.006), ischemic glomeruli Chronic CsA or FK506 nephrotoxicity: This was diag- (r = 0.4338, p = 0.026), percentage of interstitial fibrosis nosed specifically on the basis of arteriolopathy, the striped (r = 0.3965, p = 0.05) and severity of arteriosclerosis (r = fibrosis here considered non-specific in view of other pos- 0.3975, p = 0.049). Negative correlations, nearly all signif- sible causes for fibrosis. It was present in 46% of biopsies, icant, were found for these same variables versus creati- primarily in afferent arterioles, but rarely in efferent arteri- nine clearance. These correlations were yet tighter when oles, consisting of hyaline masses in the media represent- only cases with HCV were considered, for example, SCr ing necrotic myocytes. and FSGS (p = 0.7660, p = 0.001).

Diabetic nephropathy: This was present in 9 patients The importance of arteriolopathy in progression of disease (34%). Two patients had nodular glomerulosclerosis (Fig- is reﬂected in percentage counts of normal arterioles ver- ure 2D), with the typical associated vascular and tubu- sus those with calcineurin inhibitor lesions or hyaline arte- lointerstitial lesions. The remaining 7 patients had diffuse riolosclerosis. For the series as a whole, the percentage glomerulosclerosis with mesangial hypertrophy in 6 of of normal arterioles correlated well with creatinine clear- them. Typical linear IF staining was found in all. ance (p = 0.4827, p = 0.012). Of the two types of lesions, the calcineurin inhibitor arteriolopathy would appear to be Focal segmental glomerulosclerosis: FSGS was pre- more important. Considering only those patients with HCV, sent in 9 patients, particularly prominent in four biopsies, there was a strong negative correlation between creati- always in biopsies with 70% or greater interstitial ﬁbrosis nine clearance and percentage of vessels with calcineurin and particularly severe vascular lesions. inhibitor toxicity (r = 0.5760, p = 0.020), whereas the

1124 American Journal of Transplantation 2005; 5: 1120–1129 Orthotopic Liver Transplantation

2.4

2.0

1.6

1.2 Proteinuria (g/24 h) 0.8

0.4

Figure 3: Proteinuria (g/day) of patients with no diabetes nor 0.0 hepatitis C, diabetes only, hep- No HCV No Diab Diab only HCV only HCV + Diab Interferon atitis C only, diabetes and hepatitis C and of patients treated n = 7 n = 3 n = 7 n = 9 n = 7 with a -INF. correlation with hyaline arteriolosclerosis was not signifi- 0.012) and these levels were higher yet among patients cant (r = 0.1677, p = 0.53). However, specific calcineurin receiving interferon (p = 0.0022), all of whom were pro- inhibitor arteriolopathy was not correlated with either the teinuric compared with only 6 of the remaining 19 patients dose of CsA or FK506 nor to the trough levels. (Figure 3). SCrs were also worse in those treated with IFN than those not (263 ± 91 vs. 196 ± 64 lmol/L, p = Proteinuria was confined primarily to those with HCV infec- 0.047), although there was no significant difference be- tion, with higher levels in those who were simultaneously tween SCrs at the time of transplantation. They also had diabetic, and reaching its highest levels in patients treated worse glomerular lesions of all types: slcerotic glomeruli, with interferon (Figure 3). The 3 patients with diabetes only diabetic glomerulosclerosis (among the diabetics in each had only minimal proteinuria (0.03 ± 0.05 g/24 h) at the time group) and more extensive lesions of FSGS, though these of biopsy. Proteinuria is not significantly higher in patients differences were only of marginal significance individually with frank diabetic lesions or with FSGS. It appears that due to small sample size. Interestingly, there were no rec- analysis of proteinuria or urinary sediment cannot distin- ognizable differences in the extent of arteriolopathy, of in- guish those patients. terstitial fibrosis, nor of the percentage of biopsies with lesions of TMA (despite the fact that TMA has been reported Moreover, in addition to hypertension and calcineurin in- in association with interferon therapy). However, viewed in hibitor toxicity, there are two other major threats to the terms of severity, lesions of TMA were worse among the kidney, HCV infection (16 cases) and diabetes (12 cases), IFN group (chi square = 14.70, p < 0.005). These differ- which often occurred in combination (9 cases). Seven of ences are not attributable to differences in renal function 12 diabetics and 6 of 16 patients with hepatitis C went at the time of transplant. to ESRD, with 4 of these patients having both diseases. The two diseases appeared to act synergistically when in The presence of tubulointerstitial lesions related to hydrox- combination. As seen in Figure 4 when HCV and diabetes yethylstarch was not correlated with the quantity of col- occurred together, they had more severe diabetic glomer- loids received during surgery. usclerosis (p < 0.05). Factors predictive of CRF in hepatic transplantation Because of recrudescence of hepatitis in the transplant, Obviously, all of those arriving at ESRD (12 patients) were 7 patients were treated with a-IFN. These patients were hypertensive; 7 had diabetes, 6 had HCV and 4 had both compared with those with HCV who did not receive in- HCV and diabetes. The importance of calcineurin inhibitor terferon and patients without HCV. Patients with HCV had arteriolopathy is seen in the fact that 11 of 12 patients had higher levels of proteinuria than those without HCV (p = CsA/FK506 arteriolopathy and that degrees of CsA/FK506

American Journal of Transplantation 2005; 5: 1120–1129 1125 Pillebout et al.

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Diabetic glomerulosclerosis score 0.2 0.0 Diabetes - + - + + Figure 4: Diabetic glomeruloscle- VHC hepatitis - - + + + rosis score in renal biopsy of pa- α-INF treatment - - - - + tients with only diabetes, diabetes 5 and hepatitis C and diabetes and N patients 7 3 5 4 hepatitis C treated with a-IFN. arteriolopathy are associated with increasing frequency of tions, because some may be lessened or avoided entirely ESRD (grade 0 is associated with 17% of ESRD and grade by modifying the patient’s treatment. >2 with 67%). Multiple regression analysis of variables associated with ESRD at biopsy underlines the importance of HCV infection and particularly of a-IFN treatment in the Long-term consequences of CsA/FK 506 progression to ESRD (r = 0.977, p = 0.0089). The associa- nephrotoxicity tion with SCr at the time of biopsy is only to be expected. Most authors studying hepatic transplant patients with re- Because hypertension is universal (25/26 patients), as well nal failure have attributed the deterioration of renal func- as the use of calcineurin inhibitor, those factors cannot be tion to CsA/FK506 toxicity. And in fact, it has been clearly tested for by multiple regression. Table 2 summarizes his- demonstrated in patients with heart and kidney trans- tologic lesions, related to clinical data in each patient. plants (12–16,28) that immunosuppression with these drugs leads to renal vascular toxicity. This toxicity is also ob- served in hepatic transplant patients, despite substantially lower doses. Most of the hepatic transplant studies have Discussion focused on peri-operative renal function, where calcineurin inhibitors diminish the glomerular filtration rate and lead to The purpose of this article is to demonstrate the renal his- ARF (1,11,28–30). Studies of long-term renal function are tologic lesions that may be seen in hepatic transplant pa- scarce (3,9,10,28). O’Grady et al. (9) showed, at 4 years tients with chronic renal insufficiency. These lesions are post-transplant, a significant diminution in 5 of 12 patients multiple and interrelated: vascular lesions with arterioscle- studied. These patients had high doses of cyclosporine and rosis and hyaline arteriolosclerosis in almost all patients, 1 was diabetic. Another study (3), involving 120 patients calcineurin inhibitor arteriolopathy, extensive interstitial fi- at 5 years post-hepatic transplant, reported as risk factors brosis, tubular and interstitial foam cells associated with for CRF either diabetes, consumption of alcohol or both. Elhoes®, thrombotic microangiopathy and glomerular le- The authors were in agreement that in these patients, re- sions of ischemia, diabetic glomerulosclerosis and FSGS. nal function is poorly reflected by the SCr and that a renal Some of these lesions likely antedated the transplanta- biopsy would be more informative (28). Finally, a recent tion, particularly those of hypertension and diabetes, but registry analysis of 36 849 hepatic transplant patients esti- in a number of patients these conditions clearly super- mates the incidence of CRF to be 18% at 60 months and vened after transplantation. Other lesions are related to the found the following risk factors: presence of pre-operative treatment for hepatic transplantation, such as calcineurin renal failure, occurrence of post-operative ARF, the pres- inhibitor arteriolopathy tubulointerstial lesions associated ence of diabetes or hypertension, age, female sex and with Elhoes® and TMA, all likely worsened by the drugs hepatitis C infection (4). In our study as well, it appears that and the associated hypertension, particularly vascular le- diabetes, hypertension and hepatitis C infection favor pro- sions and FSGS. These have major therapeutic implica- gression toward ESRD. The nature of our study does not

1126 American Journal of Transplantation 2005; 5: 1120–1129 Orthotopic Liver Transplantation

Table 2: Histologic lesions related to some clinical data per patients Creatinine (lmo/L)/clairance (mL/min) (delay OLT/time point) Liver Before Post-operative Renal biopsy End of follow-up Dialyzed Calcineurin Histologic disease OLT (days) (years) (years) patients inhibitor a-IFN lesions HBV HCV 118/54 192/33 (13) 276/20 (7.9) 450/5 (8.9) X FK-506 Vx-HES-TMA-Ntox-FSGS HBV alcohol 80/82 166/40 (14) 281/23 (11.6) 381/5 (12.5) X CsA Vx-Ntox HCV alcohol 74/116 185/46 (15) 331/26 (3.5) 843/5 (8.0) X FK-506 X HES-FSGS HCV 81/65 161/33 (17) 154/34 (11.1) 354/5 (12.9) X CsA Vx-TMA-Ntox HCV 301/24 (5.7) 431/5 (5.9) X CsA Vx-TMA-Ntox-FSGS Alcohol 114/61 344/20 (40) 233/30 (1.0) 533/5 (2.1) X FK-506 Vx-HES-TMA-Ntox-Db HBV HCV 317/20 (11.0) 487/5 (12.1) X CsA Vx-HES-Db-FSGS HCV 89/89 178/40 (10) 129/61 (4.9) 129/61 (4.9) CsA HBV alcohol 91/74 126/53 (9) 222/30 (0.6) 188/40 (2.0) CsA HES Alcohol 93/69 132/48 (13) 180/35 (3.2) 388/5 (11.2) X FK-506 Vx-HES-TMA-Ntox-FSGS HCV 73/94 128/54 (9) 120/57 (1.1) 120/57 (1.1) FK-506 HES-TMA-Ntox HCV alcohol 140/51 370/5 (14) 182/40 (2.9) 520/5 (5.0) X CsA X HES-Db-FSGS Alcohol 79/98 711/5 (23) 168/46 (2.1) 197/44 (4.1) CsA HES-Ntox-Db Alcohol 78/76 94/63 (4) 156/38 (0.8) 145/39 (2.8) CsA Vx-HES-Ntox Alcohol 84/90 63/120 (13) 279/27 (2.2) 348/5 (3.2) X CsA Vx-HES-TMA HCV alcohol 82/99 150/54 (11) 150/64 (5.1) 159/60 (7.2) FK-506 Vx-HES-TMA-Db-FSGS HCV alcohol 111/51 139/41 (28) 445/16 (11.0) 445/28 (12.7) CsA X Vx-Db-FGS HCV 91/69 113 /56 (15) 158/41 (7.8) 144/45 (8.3) FK-506 Vx-TMA HCV 71/130 148/62 (14) 170/54 (3.5) 125/74 (4.0) FK-506 TMA HCV 86/78 198/34 (5) 146/46 (1.2) 125/54 (2.5) FK-506 X Vx-Db HCV 129/46 122/49 (21) 147/45 (4.7) 150/44 (6.1) FK-506 Vx-HES HCV 77/78 200/30 (6) 630/5 (1.3) 530/5 (2.5) X CsA X Db HCV alcohol 86/95 225/36 (25) 212/43 (5.0) 221/41 (6.0) CsA X HES-TMA-Ntox Alcohol 80/76 94/64 (7) 132/69 (9.7) 209/43 (10.5) CsA Vx-HES-TMA-Ntox-FSGS Alcohol 65/75 108/45 (4) 161/34 (3.5) 168/33 (4.8) CsA Vx-HES-Ntox HBV 100/30 114/27 (10) 265/12 (2.3) 218/15 (3.0) X FK-506 X Vx-TMA Histologic lesions = Vx severe vascular lesions; HES = hydroxylethylstarch; TMA = thrombotic microangiopathy; Ntox = nephrotoxicity of calcineurin inhibitors; Db = diabetes; FSGS = focal and segmental glomerulosclerosis; HB/CV = hepatitis B or C. Dialysed patients and patients treated by a-1FN are indicated with an X. allow to test whether the type and dosage of calcineurin In cirrhotics the incidence of glucose intolerance is esti- inhibitors are risk factors for renal deterioration. mated at 25% and insulin-dependent diabetes in another 32% (34). The problems of glucose regulation are thought In a study by Fisher et al. (31) of 883 hepatic transplants, to be secondary to muscle resistance to insulin and an in- 25 patients developed CRF by 6 month post-transplant and adequate response of the pancreatic beta cells, unable to 13 of these had a renal biopsy that showed lesions of adapt secretion of insulin to the increased peripheral de- CsA/FK506 nephrotoxicity in 10 patients and TMA in 2 pa- mand (35–38). The frequency of this association is more tients. A recent review of the literature (32) has underlined marked among patients whose cirrhosis is secondary to the frequency of renal TMA in the context of calcineurin in- hepatitis C as in our study, but without a clear patho- hibitor therapy and stresses a possible superadded genetic physiologic explanation (22,39,40). There is only a single anomaly in Factor H, proteases, etc. study dealing with the clinical consequences of this type of diabetes, particularly its vascular complications (35). It TMA and a-IFN in hepatic transplantation compares 34 cirrhotic diabetics to diabetics without liver TMA during a-IFN treatment occurs most frequently in disease. The cirrhotic diabetics had a lower incidence of hematologic disorders (17). It is only recently that cases diabetic retinopathy but greater proteinuria. This study un- of renal (20,21) and/or cutaneous TMA (33) have been re- fortunately did not deal with renal function. The diabetic ported during treatment of hepatitis C by a-IFN. In our pa- nephropathy in our patients appeared severe and rapidly tients we suggest that the association of a-IFN and cal- evolving, probably aggravated by calcineurin inhibitors. In cineurin inhibitors has augmented the severity of TMA. effect, it has been demonstrated that cyclosporine and to a greater extent tacrolimus induce modiﬁcations of in- Diabetic nephropathy and hepatic transplantation sulin secretion leading to diabetes, involving up to 36% of The incidence of diabetes in our series (47%) is far transplant patients treated with tacrolimus, perhaps ex- higher than in the general population. The development plaining the de novo diabetes post-liver-transplant (41–43). of cirrhosis-associated impaired glucose tolerance, if not In our study it was not possible to implicate a speciﬁc cal- frank diabetes ‘of hepatic origin’, is well recognized (22,34– cineurin inhibitor in the development of diabetic nephropa- 36). thy. The diabetes of these cirrhotic-then-transplanted

American Journal of Transplantation 2005; 5: 1120–1129 1127 Pillebout et al. patients appears thus to be different from the type II di- plantation until it can be definitely proven that it does not abetes seen in the general population. It appears to have play a role in ARF and CRF in hepatic transplantation. It a specific pathophysiologic mechanism and fewer extra- would also be preferable, as other authors recommend renal vascular complications, contrasting with the rapidly (29) not to begin cyclosporine or tacrolimus until 72 h evolving renal diabetic lesions. post-operatively when renal hemodynamic have returned toward normal. Chronic hydroxyethylstarch nephrotoxicity We were surprised to find as long as 10 years after the use Finally, when hepatitis C viral replication menaces the func- of hydroxyethylstarch (Elhoes®), the persistence of the his- tion of the hepatic graft and it is necessary to institute tologic lesions previously described, but only in the months treatment with a-IFN, this must be done with caution, with following its use. The presence of these tubulointerstitial careful surveillance of renal function and signs of hemol- lesions in 61% of the native kidneys of our patients raises ysis, and possible institution of antiplatelet-aggregating the question of its possible interstitial toxicity, particularly agents. in view of the greater tubular atrophy in these patients. Of course, the attribution of the tubular and interstitial foam cells to HES overload is strictly one of association. They References are present in patients who have received this product, and absent in those who have not. This, plus early stud- 1. Platz KP, Mueller AR, Blumhardt G et al. Nephrotoxicity follow- ies showing these cells in a more acute setting where the ing orthotopic liver transplantation. A comparison between cy- time course and dosages were available, leads us to the closporine and FK506. Transplantation 1994; 58: 170–178. tentative conclusion that HES is the cause of these lesions. 2. Gonwa TA, Klintmalm GB, Levy M, Jennings LS, Goldstein RM, Three studies have, in fact, demonstrated the renal tox- Husberg BS. Impact of pretransplant renal function on survival icity of this colloid, responsible for ARF shortly after its after liver transplantation. Transplantation 1995; 59: 361–365. 3. Gayowski T, Singh N, Keyes L et al. Late-onset renal failure after utilization and have described the characteristic histologic liver transplantation: role of posttransplant alcohol use. Transplan- lesions (26,44,45). In these three studies the renal failure tation 2000; 69: 383–388. was acute and reversible. No study to date has shown 4. Ojo AO, Held PJ, Port FK et al. Chronic renal failure after trans- the chronic nephrotoxicity of this product. These ‘osmotic’ plantation of a nonrenal organ. N Engl J Med 2003; 349: 931–940. lesions seem likely to represent an intra-cytoplasmic over- 5. Bataller R, Gines P, Guevara M, Arroyo V. Hepatorenal syndrome. load of the colloid or its metabolites. The final proof will Semin Liver Dis 1997; 17: 233–247. be the identification of the material in the cells by electron 6. Epstein M. Hepatorenal syndrome: emerging perspectives. microscopy or other techniques since it does not polarize. Semin Nephrol 1997; 17: 563–575. rd However, the association is too strong to be ignored. 7. Skluzacek PA, Szewc RG, Nolan CR, 3 , Riley DJ, Lee S, Pergola PE. Prediction of GFR in liver transplant candidates. Am J Kidney Dis 2003; 42: 1169–1176. Therapeutic consequences 8. Hiesse C, Samuel D, Chraibi A et al. Combined liver and kid- It is therefore necessary to manage these hepatic trans- ney transplantation in nephropathy associated with liver disease. plant patients at great risk of CRF in global terms, taking all Transplant Proc 1995; 27: 1262–1263. of the factors into consideration. 9. O’Grady JG, Forbes A, Rolles K, Calne RY, Williams R. An analysis of cyclosporine efficacy and toxicity after liver transplantation. Diabetes, proven histologically by our study, but often un- Transplantation 1988; 45: 575–579. recognized or ignored, seems not to have been sufficiently 10. Iwatsuki S, Starzl TE, Todo S et al. Experience in 1000 liver trans- taken into account in patient care. Careful surveillance and plants under cyclosporine-steroid therapy: a survival report. Trans- optimal treatment are necessary. The absence of retinopa- plant Proc 1988; 20(1 Suppl. 1): 498–504. thy or other vascular complications should not reassure the 11. Platz KP, Mueller AR, Blumhardt G et al. Nephrotoxicity after orthotopic liver transplantation in cyclosporin A and FK 506-treated clinician. It has now been demonstrated that optimal treat- patients. Transpl Int 1994; 7(Suppl. 1): S52–S57. ment of diabetes considerably diminishes the risk of devel- 12. Myers BD, Ross J, Newton L, Luetscher J, Perlroth M. oping diabetic nephropathy (46,47). Finally, identification of Cyclosporine-associated chronic nephropathy. N Engl J Med microalbuminuria should lead to institution of angiotensin 1984; 311: 699–705. converting enzyme treatment, which has also been shown 13. Myers BD, Newton L, Boshkos C et al. Chronic injury of human re- to slow the progression of diabetic nephropathy (48–51). nal microvessels with low-dose cyclosporine therapy. Transplan- tation 1988; 46: 694–703. Moreover, the majority of our patients were hypertensive 14. Bennett WM, DeMattos A, Meyer MM, Andoh T, Barry JM. but without adequate anti-hypertensive treatment. Optimal Chronic cyclosporine nephropathy: the Achilles’ heel of immuno- anti-hypertensive therapy is also necessary in order to slow suppressive therapy. Kidney Int 1996; 50: 1089–1100. 15. Andoh TF, Burdmann EA, Bennett WM. Nephrotoxicity of im- the progression of CRF, as has been clearly demonstrated munosuppressive drugs: experimental and clinical observations. (52,53). Semin Nephrol 1997; 17: 34–45. 16. Franceschini N, Alpers CE, Bennett WM, Andoh TF. Cyclosporine We would also suggest the avoidance of the use of hydrox- arteriolopathy: effects of drug withdrawal. Am J Kidney Dis 1998; yethylstarch as an intra-operative colloid in hepatic trans- 32: 247–253.

1128 American Journal of Transplantation 2005; 5: 1120–1129 Orthotopic Liver Transplantation

17. Zuber J, Martinez F, Droz D, Oksenhendler E, Legendre C. 38. Perseghin G, Mazzaferro V, Sereni LP et al. Contribution of re- Alpha-interferon-associated thrombotic microangiopathy: a clin- duced insulin sensitivity and secretion to the pathogenesis of icopathologic study of 8 patients and review of the literature. hepatogenous diabetes: effect of liver transplantation. Hepatol- Medicine (Baltimore) 2002; 81: 321–331. ogy 2000; 31: 694–703. 18. Badid C, McGregor B, Faivre JM et al. Renal thrombotic mi- 39. Caronia S, Taylor K, Pagliaro L et al. Further evidence for an as- croangiopathy induced by interferon-alpha. Nephrol Dial Trans- sociation between non-insulin-dependent diabetes mellitus and plant 2001; 16: 846–848. chronic hepatitis C virus infection. Hepatology 1999; 30: 1059– 19. Guerret S, Desmouliere A, Chossegros P et al. Long-term ad- 1063. ministration of interferon-alpha in non-responder patients with 40. Mehta SH, Brancati FL, Sulkowski MS, Strathdee SA, Szklo M, chronic hepatitis C: follow-up of liver fibrosis over 5 years. J Viral Thomas DL. Prevalence of type 2 diabetes mellitus among per- Hepat 1999; 6: 125–133. sons with hepatitis C virus infection in the United States. Hepa- 20. Iyoda K, Kato M, Nakagawa T et al. Thrombotic thrombocy- tology 2001; 33: 1554. topenic purpura developed suddenly during interferon treatment 41. Mayer AD, Dmitrewski J, Squifflet JP et al. Multicenter random- for chronic hepatitis C. J Gastroenterol 1998; 33: 588–592. ized trial comparing tacrolimus (FK506) and cyclosporine in the 21. Ubara Y, Hara S, Takedatu H et al. Hemolytic uremic syndrome prevention of renal allograft rejection: a report of the European associated with beta-interferon therapy for chronic hepatitis C. Tacrolimus Multicenter Renal Study Group. Transplantation 1997; Nephron 1998; 80: 107–108. 64: 436–443. 22. Naunyn B. Der Diabete Mellitus. Wien, Austria: Holder, 1898: 38. 42. Knoll GA, Bell RC. Tacrolimus versus cyclosporin for immunosup- 23. Nochy D, Hill GS, Antoine C, Glotz D, Calmus Y. Diabetic pression in renal transplantation: meta-analysis of randomised tri- nephropathy—a frequent complication of liver transplant? J Am als. Br Med J 1999; 318: 1104–1107. Soc Nephrol 2000; 11: A3805. 43. Jain AB, Kashyap R, Rakela J, Starzl TE, Fung JJ. Primary adult 24. Nankivell BJ, Borrows RJ, Fung CL, O’Connell PJ, Chapman JR, liver transplantation under tacrolimus: more than 90 months ac- Allen RD. Calcineurin inhibitor nephrotoxicity: longitudinal assess- tual follow-up survival and adverse events. Liver Transpl Surg ment by protocol histology. Transplantation 2004; 78: 557–565. 1999; 5: 144–150. 25. McCauley J. The nephrotoxicity of FK506 as compared with cy- 44. Cittanova ML, Leblanc I, Legendre C, Mouquet C, Riou B, Coriat closporine. Curr Opin Nephrol Hypertens 1993; 2: 662–669. P. Effect of hydroxyethylstarch in brain-dead kidney donors on 26. Legendre C, Thervet E, Page B, Percheron A, Noel LH, Kreis H. renal function in kidney-transplant recipients. Lancet 1996; 348: Hydroxyethylstarch and osmotic-nephrosis-like lesions in kidney 1620–1622. transplantation. Lancet 1993; 342: 248–249. 45. Schortgen F, Lacherade JC, Bruneel F et al. Effects of hydrox- 27. Nochy D, Daugas E, Droz D et al. The intrarenal vascular lesions yethylstarch and gelatin on renal function in severe sepsis: a mul- associated with primary antiphospholipid syndrome. J Am Soc ticentre randomised study. Lancet 2001; 357: 911–916. Nephrol 1999; 10: 507–518. 46. The Diabetes Control and Complications Trial Research Group. 28. de Mattos AM, Olyaei AJ, Bennett WM. Nephrotoxicity of The effect of intensive treatment of diabetes on the development immunosuppressive drugs: long-term consequences and chal- and progression of long-term complications in insulin-dependent lenges for the future. Am J Kidney Dis 2000; 35: 333–346. diabetes mellitus. N Engl J Med 1993; 329: 977–986. 29. Distant DA, Gonwa TA. The kidney in liver transplantation. J Am 47. The Diabetes Control and Complications (DCCT) Research Group. Soc Nephrol 1993; 4: 129–136. Effect of intensive therapy on the development and progression 30. Fraley DS, Burr R, Bernardini J, Angus D, Kramer DJ, Johnson JP. of diabetic nephropathy in the Diabetes Control and Complica- Impact of acute renal failure on mortality in end-stage liver disease tions Trial. Kidney Int 1995; 47: 1703–1720. with or without transplantation. Kidney Int 1998; 54: 518–524. 48. Hallab M, Gallois Y, Chatellier G, Rohmer V, Fressinaud P, Marre 31. Fisher NC, Nightingale PG, Gunson BK, Lipkin GW, Neuberger M. Comparison of reduction in microalbuminuria by enalapril and JM. Chronic renal failure following liver transplantation: a retro- hydrochlorothiazide in normotensive patients with insulin depen- spective analysis. Transplantation 1998; 66: 59–66. dent diabetes. Br Med J 1993; 306: 175–182. 32. Schwimmer J, Nadasdy TA, Spitalnik PF, Kaplan KL, Zand MS. De 49. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD. The effect of novo thrombotic microangiopathy in renal transplant recipients: a angiotensin-converting-enzyme inhibition on diabetic nephropa- comparison of hemolytic uremic syndrome with localized renal thy. The Collaborative Study Group. N Engl J Med 1993; 329: thrombotic microangiopathy. Am J Kidney Dis 2003; 41: 471– 1456–1462. 479. 50. UKPDS 38. UK Prospective Diabetes Study Group. Tight blood 33. Creput C, Auffret N, Samuel D, Jian R, Hill G, Nochy D. Cuta- pressure control and risk of macrovascular and microvascular neous thrombotic microangiopathy during treatment with alpha- complications in type 2 diabetes. Be Med J 1998; 317: 703– interferon for chronic hepatitis C. J Hepatol 2002; 37: 871– 713. 872. 51. UKPDS 39. UK Prospective Diabetes Study Group. Efficacy of 34. Megyesi C, Samols E, Marks V. Glucose tolerance and diabetes atenolol and captopril in reducing risk of macrovascular and mi- in chronic liver disease. Lancet 1967; 2: 1051–1056. crovascular complications in type 2 diabetes. Br Med J 1998; 317: 35. Vidal J, Ferrer JP, Esmatjes E et al. Diabetes mellitus in patients 713–720. with liver cirrhosis. Diabetes Res Clin Pract 1994; 25: 19–25. 52. Peterson JC, Adler S, Burkart JM et al. Blood pressure control, 36. Kruszynska YT, Goulas S, Wollen N, McIntyre N. Insulin secretory proteinuria, and the progression of renal disease. The Modifica- capacity and the regulation of glucagon secretion in diabetic and tion of Diet in Renal Disease Study. Ann Intern Med 1995; 123: non-diabetic alcoholic cirrhotic patients. J Hepatol 1998; 28: 280– 754–762. 291. 53. Lazarus JM, Bourgoignie JJ, Buckalew VM et al. Achievement 37. Petrides AS, Schulze-Berge D, Vogt C, Matthews DE, Strohmeyer and safety of a low blood pressure goal in chronic renal disease. G. Glucose resistance contributes to diabetes mellitus in cirrhosis. The Modification of Diet in Renal Disease Study Group. Hyper- Hepatology 1993; 18: 284–291. tension 1997; 29: 641–650.

American Journal of Transplantation 2005; 5: 1120–1129 1129