Early Presence of Calcium Oxalate Deposition in Kidney Graft Biopsies is Associated with Poor Long-Term Graft Survival

Authors


*Corresponding author: Alvaro Pacheco-Silva, apacheco@nefro.epm.br

Abstract

Accumulated oxalate will be excreted after renal transplantation, creating an increased risk of tubular precipitation, especially in the presence of allograft dysfunction. We evaluated calcium oxalate (CaOx) deposition in renal allograft biopsies with early dysfunction, its association with acute tubular necrosis (ATN) and graft survival. We studied 97 renal transplant patients, submitted to a graft biopsy within 3 months post-transplant, and reanalyzed them after 10 years. We analyzed renal tissue under polarized light and quantified CaOx deposits. CaOx deposits were detected in 52.6% of the patients; 26.8% were of mild and 25.8% of moderate intensity. The deposits were more frequent in biopsies performed within 3 weeks post-transplant (82.4 vs. 63.0%, p < 0.05) and in allografts with more severe renal dysfunction (creatinine 5.6 mg/dL vs. 3.4 mg/dL, p < 0.001). ATN incidence was also higher in patients with CaOx deposits (47% vs. 24%, p < 0.001). Twelve-year graft survival was strikingly worse in patients with CaOx deposits compared to those free of deposits (49.7 vs. 74.1%, p = 0.013). Our study shows a high incidence of CaOx deposits in kidney allografts with early dysfunction, implying an additional risk for acute tubular injury, with a negative impact on graft survival.

Introduction

Oxalate (Ox) is an end-product of many metabolic pathways. Thus, it cannot be further processed, being totally eliminated through glomerular filtration by the kidneys (1,2). In patients with chronic renal disease (CRD), Ox accumulates 10–30 times above normal levels as a result of its reduced excretion (2–6). Neither hemodialysis nor peritoneal dialysis is able to normalize Ox levels in CRD patients. Only a 60% reduction of Ox pool is expected to happen after a usual hemodialysis procedure, but Ox was found to return to pre-dialysis levels within 48 h (7). In contrast to primary hyperoxalurias, clinical manifestations of uremic oxalosis, such as nephrolithiasis, fractures and bone pain, are not common (2,8,9).

After an immediately successful renal transplantation, the graft will rapidly excrete the accumulated Ox, with its serum levels reaching the normal range in a period of 5 days to 3 weeks. This tremendous excretion of calcium oxalate (CaOx), within a short time, generates conditions for renal tissue deposition and damage, especially in favorable conditions (5,6). CaOx deposition in renal tissue after transplantation in patients without primary hyperoxaluria has been poorly studied, especially if there is concomitant allograft dysfunction such as acute tubular necrosis (ATN) (10,11). Few authors have observed CaOx deposits in post-transplant ATN, and considering them as an additional cause for tubular cell injury, their correlation with ATN and allograft long-term survival is a matter of interest.

Recently, many transplantation research groups have focused their attention to non-acute rejection early events that might interfere with graft survival, mainly delayed graft function. Although the incidence of acute rejection has been decreasing steadily with the new era of immunosuppressive drugs, the beneficial effects of this reduction on graft outcomes has not been so clear (12). Some authors have suggested that the renal function at 1st year post-transplantation and the incidence of delayed graft function were responsible for this lack of improvement on graft survival (13,14).

Hence, we were interested in evaluating the incidence of CaOx deposits in renal allografts with early dysfunction, and in correlating them with graft outcomes.

Methods

Patients

We studied all patients who received a renal transplant between August 1990 and August 1994 at Universidade Federal de São Paulo, a tertiary reference hospital center in São Paulo, and who were submitted to at least one allograft biopsy due to renal dysfunction during the first 3 months post-transplantation, regardless of donor type. Renal dysfunction was defined by an increase of 30% on baseline serum creatinine levels or a sustained graft non-function. The total number of transplant patients was 346 and an allograft biopsy was performed in 151 of them (44%). We excluded 54 individuals for the following reasons: a biopsy specimen was not available (N = 34), allograft loss due to vascular thrombosis, with necrotic tissue making difficult oxalate analysis (N = 13), a history of kidney stones (N = 6) and a previous diagnosis of primary hyperoxaluria (N = 1). Thus, 97 patients were initially studied in 1994 and then, these patients were reanalyzed after 10 years of follow-up. A total of 175 biopsy specimens were analyzed, median of 1.0 biopsy per patient (range, 1–7); 52.6% were submitted to one allograft biopsy and 78.4% had one or two biopsies. Acute rejection (60%) and acute tubular necrosis (36%) were the main histological findings. Patient data were obtained from the hospital records. The Ethical Committee on Human Studies of our institution approved this study.

Transplant and clinical variables

Transplant surgery was performed according to normal surgical and anaesthesic procedures. Methoxyflurane an anaesthesic that may cause intense Ox production, was avoided (15). All patients received a triple immunosuppression regimen with prednisone, azathioprine and cyclosporine A. We assessed gender, donor and recipient ages, etiology and duration of CRD, type and duration of dialysis, donor type and early clinical graft outcomes, such as total daily urine output at the first day after engraftment and serum creatinine levels at the time of biopsies.

Post-transplant acute tubular necrosis (ATN) was defined as the requirement for dialysis within the first week after transplant, excluding acute rejection and technical causes of renal dysfunction, even without histological evaluation. ATN duration was the period between the first and the last dialysis procedure after transplantation surgery. We computed cold ischemia, warm ischemia and anastomosis times.

The diagnosis of acute rejection was based on clinical data and in most cases it was confirmed by histological criteria (60%, 50/83). The rejection episodes were treated with pulse therapy with methylprednisolone (1 g/day, 3–5 days). Seventy-five percent of patients had cellular acute rejection and, OKT3 was used in 30% of all acute rejection cases.

Graft loss was defined by the requirement of permanent dialysis after graft failure, and death of patients with functioning grafts was not considered an end-point event.

Histopathological analysis

Percutaneous renal transplant biopsies were performed and processed routinely for light microscopy after fixation using Bouin's fixative and paraffin embedding. Hematoxylin and eosin stained sections were examined under polarized light with a microscope at 400× by one pathologist who was blinded to patient information. The CaOx deposition was quantified by the score proposed by Scheinman et al. (16). Briefly, full, non-overlapping cortical fields were examined and the number of cross sections containing oxalate crystals was counted and scored. A score of 3 was assigned when crystals occupied the entire tubule, a score of 1 when a crystal filled no more than 5–10% of the tubule lumen, and 2, when an intermediate amount of oxalate crystals was observed. The biopsy score is reported as the total number of oxalate-containing tubule cross sections times the crystals scores, divided by the number of cortical fields studied (5–20 fields). According to the biopsy score, the intensity of CaOx deposition was classified as mild (a score below 0.7), moderate (a score between 0.7 and 7.0) and severe, when the score was higher than 7.0.

Statistical analysis

Comparisons of continuous variables were performed by the Student's t-test, or the Mann-Whitney U-test where appropriate; and the chi-square or Fisher's exact tests were used to compare categorical variables. Kaplan-Meier analysis was used to calculate allograft and patient survival rates and the log-rank test was employed to compare allograft and patient survival curves. Univariate analyses were performed with available data related to graft survival. Those variables that had a positive correlation were further tested in multivariate analyses. To investigate whether there was an independent association between graft failure and Ox deposition, a logistic regression analysis was performed using a stepwise forward model. The following independent variables were tested: CaOx deposition (yes vs. no), cold ischemia time (>20 h vs. <20 h), donor age (<34 years vs. >34 years), time on dialysis (>28 months vs. <28 months), ATN (yes or no), type of donor (deceased vs. living donor), total daily urine output at first day after engraftment (<3370 mL vs. >3370 mL), acute rejection (yes vs. no) and serum creatinine levels at biopsy time (<4.8 mg/dL vs. >4.8 mg/dL). We examined the proportional hazard assumption by plotting the graft curves for each group of a covariate on a log-log scale. Because the curves appeared reasonably parallel, we regarded the model as appropriate. Adjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated.

Results are expressed as the mean ± SD or median (minimum – maximum). All p-values were two-tailed and a value <0.05 was considered to indicate statistical significance. The analysis was done using the SPSS 10.0 Statistical Software (Chicago, IL, 1992).

Results

Patients

Ninety-seven patients (65 men and 32 women) ranging in age from 9 to 60 years (34.0 ± 12.2 years) were enrolled in this study. Ninety-one patients (93.8%) were recipients of a first transplant, 5 (5.1%) of a second and one (1.1%) of a third renal transplant. Concerning the donors, 75 (73.3%) patients received grafts from deceased donors and 22 (26.7%) from living donors (18 haplo-identical HLA, 1 identical HLA, 3 distinct HLA). The primary diagnosis of CRD was chronic glomerulonephritis in 37 (38.1%), hypertension in 15 (15.4%), chronic pyelonephritis in 9 (9.2%), polycystic kidney disease in 7 (7.2%), other causes in 3 (3.0%) and unknown in 22 (27.1%). Seventy-one patients (73.2%) were on hemodialysis and 25 on peritoneal dialysis (25.8%) and 1 did not receive any dialysis therapy (1.0%). Median time of renal insufficiency was 38.5 months (6–196 months), and mean time on dialysis was 33.3 ± 23.6 months.

Oxalate deposition

The overall incidence of CaOx deposits in the study population was 52.6% (51/97). The CaOx deposits were found only in the tubular lumen, and under no polarized light examination, they were seen as an intra-tubular gray material, hardly remembering granular casts. According to the quantitative score, 26 individuals (26.8%) had a mild deposition and 25 (25.8%) a deposition of moderate intensity (Figure 1). We did not find any patient with severe deposition (Table 1).

Figure 1.

Histological appearance of a transplant kidney biopsy examined under polarized light showing a moderate amount of calcium oxalate crystals in the tubules. H&E, 100×.

Table 1.  Clinical characteristics of patients according to presence or absence of CaOx deposits
 OXA groupWOXA groupp-values
  1. OXA = oxalate deposition, WOXA = without oxalate deposition, CRD = chronic renal disease, HD = hemodialysis.

  2. Data are expressed as median (range) or mean ± SD.

  3. *One patient in WOXA did not receive any renal replacement therapy.

%, N52.6 (51/97)47.4 (46/97)
Intensity of deposition
 Mild51.0 (26/51) 
 Moderate49.0 (25/51) 
Time of biopsy (days)11.0 (1.0–70.0)10.0 (1.0–87.0)0.308
Time of biopsy, (<3 weeks, %, N)84.4 (42/51)63.0 (29/46)<0.05
Creatinine at biopsy (mg/dL)5.6 (1.5–17.0)3.4 (1.2–4.0)<0.001
Number of biopsies2.0 (1.0–7.0)1.0 (1.0–4.0)<0.05
Urine output at 1st day (mL)1440 (0–12870)5585 (0–20100)0.003
Sex, male/female (%, N)62.7/37.3 (32/19)65.2/34.8 (30/16)0.800
Recipient age (years)35.0 (9.0–57.0)37.5 (9.0–60.0)0.439
Donor age (years)34.4 ± 14.233.8 ± 15.30.809
Etiology of CRD (%, N)
 Chronic glomerulonephritis35.3 (18/51)41.3 (19/46) 
 Hypertensive17.6 (09/51)13.0 (06/46) 
 Unknown21.6 (11/51)26.1 (12/46) 
 Others25.5 (13/51)19.6 (12/46)0.767
Renal disease (months)38.0 (7.0–196.0)42.0 (6.0–168.0)0.269
Time on dialysis (months)51.0 (5.0–108.0)46.0 (0.0–102.0)0.893
Type of dialysis (%, N)
 HD74.5 (38/51)71.7 (33/46) 
 Peritoneal25.5 (13/51)26.1 (12/46)*0.566
Deceased donors (%, N)84.3 (43/51)69.6 (32/46)0.083
Acute rejection (%, N)90.2 (46/51)80.4 (37/46)0.248
Number of acute rejection1.0 (0–3)1.0 (0–4)0.370

In 86.4% of the patients the CaOx deposits were already found in the first biopsy. Regarding the time the allograft biopsy was done, the median time post-transplant in the CaOx group (OXA) was 11.0 days (1.0–70.0 days), similar to the group without CaOx deposits (WOXA) (10.0 days, range 1.0–87.0 days), p = 0.308. We observed that the majority, 82.4% (42/51), of the CaOx positive biopsies were performed within 3 weeks post-transplant, in contrast to 63% (29/46) in the WOXA group, p < 0.05 (Table 1).

Looking at renal function at the time of biopsy, the OXA group had more severe renal dysfunction since they presented higher median levels of creatinine compared to the WOXA group, 5.6 mg/dL (1.5–17.0) versus 3.4 mg/dL (1.2–15.3), respectively, p < 0.001. Indeed, patients in WOXA group were submitted to a few renal biopsies than the OXA group, median 1.0 (1.0–4.0) versus 2.0 (1.0–7.0), respectively, p < 0.05. Patients in OXA group had a low daily urine output at first day after engraftment (1440 mL vs. 5585 mL, p = 0.003), when compared to WOXA patients (Table 1).

There was no significant difference in the gender distribution, donor and recipient ages, cause of CRD, duration of renal insufficiency, duration of renal replacement therapy and type of dialysis. Deceased donors tended to be more frequent in the OXA group, 84.3% of the donors, in contrast to 69.6% in the WOXA group, although not reaching statistical significance, p = 0.08 (Table 1). The occurrence and the median number of acute rejection episodes did not differ significantly between groups (Table 1).

Acute tubular necrosis versus oxalate deposition

ATN was present in 35 patients of our study population, corresponding to an incidence of 36% (35/97). Among the 51 patients with CaOx deposits, 24 (47%) were in the ATN group. However, 24% of patients (11/51) without OXA (WOXA) presented ATN in their evolution. Although it is difficult to link ATN to renal CaOx deposits, since the latter could be implicated in ATN pathogenesis and, in another way out, ATN could predispose to CaOx crystals deposition, we investigated the risk to develop ATN in the presence of Ox deposition. There was a higher incidence of ATN among patients from OXA group when compared with patients without any deposition (47% vs. 24%, p < 0.05). The OR of having ATN in the presence of CaOx deposits was 2.83 (95% CI, 1.18–6.76, p = 0.018) (Table 2).

Table 2.  Development and duration of ATN according to the presence of CaOx deposition
 OXA group
(N = 51)
WOXA group
(N = 46)

p-values
  1. Data are presented as median and range (minimum and maximum).

  2. Odds ratio of ATN in presence of CaOx – 2.83 (95% CI, 1.18–6.76, p = 0.018).

ATN (%, N)
 Yes47.0 (24/51)24.0 (11/46) 
 No53.0 (27/51)76.0 (35/46)0.018
Time on ATN (days)15.5 (4.0–66.0)9.0 (6.0–13.0)0.001

When we considered only the patients who developed ATN (35/97), and stratified according to the presence or not of CaOx deposits, we did not observe significant difference in the patient age, gender, cause of CRD, duration of renal insufficiency, duration of renal replacement therapy and type of dialysis. Proportion of deceased donors, cold ischemia, warm ischemia and vascular anastomosis times did not also differ between groups. However, the median time of recovery from ATN, assessed by the time spent from the day of transplant to the last dialysis day, was longer in the patients from the OXA group than in the WOXA group (15.5 days [4–66] vs. 9 days [6–13], p = 0.001) (Table 2).

Graft and patient survival rates

Patient and allograft survival rates were worse in patients with CaOx deposits (OXA group). Indeed, patient survival rate in this group was 95.5% in the first year and 74.1% at 12 years. In patients without CaOx deposits, the first year survival was 98.6% and 79.3% after 12 years (p < 0.01). In the OXA group, allograft survival was 72.5% at the end of the first year and 49.7% at 12 years; the corresponding values in the WOXA group were remarkably better, 89.1% and 74.1% (p = 0.013) (Figure 2). We further investigated the relationship between CaOx alone and graft function by analyzing the graft survival in patients free of ATN that had CaOx deposition or not. Considering this stratification, we observed that CaOx deposition was implicated in worse graft survival in those selected patients free of ATN (47.9% vs. 74.9%, p = 0.0129) (Figure 3).

Figure 2.

Graft survival in patients with CaOx deposits (dashed line) was 49.1% after 12 years, compared to 74.1% in patients with no deposits (filled line), p < 0.05.

Figure 3.

Graft survival in ATN-free patients without CaOx deposits (filled line) was 74.9% after 12 years, compared to 47.9% in patients with deposits (dashed line), p < 0.05.

Univariate analyses demonstrated that serum creatinine at biopsy time and daily urine output at first day after surgery were significantly associated with CaOx crystal deposition (Table 1). However, logistic regression analysis confirmed the presence of a statistically significant and independent association between return to dialysis and oxalate deposition. Patients with oxalate deposition had 4.32 times higher risk for graft failure (95% CI, 1.182–15.821) than those in the WOXA group (p = 0.027). Any other variable tested were independently associated to graft lost (Table 3).

Table 3.  Logistic regression to graft failure

Variables

p-values
Exp
(B) = RR

CI 95%
Oxalate, Yes versus No0.0274.3241.182–15.821
Cold ischemia time (>20 h)0.0530.2780.076–1.015
Donor age (>34 years)0.1642.3220.707–7.689
Dialysis time (>28 months)0.2202.1260.637–7.086
ATN, Yes versus No0.2970.4430.096–2.046
Deceased donor0.2422.6080.524–12.969
Urine output at 1st day (>3370 mL)0.5460.5770.097–3.439
Acute rejection, Yes versus No0.5330.5730.10–3.300
Serum creatinine at biopsy time (>4.8 mg/dL)0.6150.4800.27–8.383

Discussion

Chronic renal failure is the most frequent cause of oxalosis (uremic oxalosis), in which Ox accumulates due to renal failure. The levels of Ox can achieve super saturation and precipitate as birefringent CaOx crystals preferentially in kidney, bone, joints, cardiac conductive system, blood vessels and retina (2,17–19). End-stage renal disease patients have elevated serum levels of Ox, about 10-fold the normal value (4). Even though, dialysis is not efficient in removing Ox, clinical manifestations of secondary oxalosis are rare (2,20,21).

At the time of renal transplant, the expanded Ox pool will be eliminated, sometimes under favorable conditions to its deposition, such as ATN. Worchester et al. (5) showed normalization of the urinary and the serum levels of Ox 5 days post-transplant, in living-related recipients with good graft function. More recently, Hoppe et al. showed a longer period, of 3 weeks, to normalization of Ox serum levels after a successful renal engraftment in children (6).

In order to study the incidence of CaOx deposition, we examined kidney biopsy samples performed during the first 3 months after transplantation, a period where biopsies are frequently performed to clarify the etiology of allograft dysfunction. The impact of CaOx deposition on graft outcome was then investigated.

The incidence of CaOx deposits on allograft biopsy ascertained by light microscopy was somewhat higher than what we expected (52.6%), with half of the cases being of mild intensity and half of moderate intensity. Farnsworth et al. reported a lower CaOx deposition (40%) in 56 transplant patients with ATN, but they did not evaluate the factors associated with deposition and only performed a qualitative analysis (10). Olsen et al. mentioned the presence of CaOx crystals in 77% of patients with ATN during the pre-cyclosporine era, but they did not include all causes of delayed graft function (11). Solez et al. reported Ox deposition of the same intensity in biopsies from patients with ATN and from patients with acute cyclosporine nephrotoxicity, although they did not describe the method of Ox quantification or the exact extension of its deposition (22).

We could not find any correlation among gender, donor or recipient ages, etiology of CRD, presence or number of acute rejection episodes and CaOx deposits. Interestingly, a longer duration of CRD and dialysis therapy, which might suggest a greater amount of accumulated Ox, were not significantly associated with CaOx deposition in our population. This finding is in contrast to other authors' data who reported increased CaOx deposits in native kidneys from patients with more than 3 months on dialysis (20,23); and with others who described more deposits in patients on dialysis when compared with patients in the pre-dialysis phase (2,24). We also suspected that patients on peritoneal dialysis were at increased risk of CaOx deposition, since Ox dialysance is found to be less efficient and oxalate serum levels were described to be higher in patients undergoing peritoneal dialysis (7,25). Our study did not enforce that hypothesis.

We found an interesting correlation between CaOx deposits and ATN. ATN was present in 47% of the patients with deposits, but in only 24% of the patients without CaOx deposits. On the other hand, CaOx crystals were found in 68% of the patients with ATN (24/35), a percentage slightly lower than that (77%) reported by Olsen et al. (11). Parameters closely related to ATN, such as greater cold, anastomosis and warm ischemia times, were not associated with CaOx deposition. We also observed, pointing out again to a pattern of outstanding allograft injury, a longer recovery time from ATN in patients with CaOx deposition (median of 16 vs. 11 days).

Although it is difficult to establish a cause and effect relationship between ATN and CaOx deposits, which one is in fact cause or consequence, it is nonetheless, possible to infer that the presence of crystals per se is injurious to tubular cells. For this purpose, we performed some additional analyses where we analyzed only patients free of ATN. This specific subpopulation was further stratified in two groups according to the presence or not of CaOx. We detected a more striking difference. ATN-free patients with CaOx had a worse graft survival than ATN-free patients without CaOx (47.9% vs. 74.7%, p = 0.0129).

The presence of Ox deposits was associated with a poorer graft survival. The presence of CaOx deposits implied a decrease of almost 30% in graft survival. Logistic regression confirmedOx deposition as a condition independently associated to graft loss. Patients with CaOx deposits had a 4.32-fold increased risk of losing their grafts, independently of other variables, compared to those free of deposits.

Besides causing obstructive damage, Ox is implicated in direct injury to tubular cells. Some authors reported increased urinary excretion of tubular enzymes in experimental studies of induced hyperoxaluria (26–28) and in CaOx stone formers clinical setting (29). Tubular cells death occurs by oxygen-free radicals-mediated necrosis, dependent on Ox concentration (30) or by apoptosis (31). Ox seems to exert a biphasic effect on tubular cells, which is toxic at high concentrations but acts as a mitogen at lower levels (32). It have been also demonstrated that an increase in transcription of activating factors (c-myc, EGR-1, Nur-77, c-jun), MAP kinases, extracellular matrix regulators and growth factors, suggests an involvement of Ox in fibrogenesis of the renal interstitium (33–35). In this sense, the presence of intra-luminal crystals may induce tubular cells to secrete pro-fibrotic factors and promote renal scars.

The most accepted theory of chronic allograft nephropathy (CAN) states that it emerges in transplants that have undergone previous damage, from immunologic to non-immunologic causes. Following the damage, an impaired tubular repair leads to the characteristic excessive fibrosis and loss of function (36–38). Recently, Nankivell et al. had strongly supported this theory, in a clinical study based on protocol biopsies. They found that early allograft injury by ischemic injury, severe rejection and subclinical rejection predicted CAN (39). Here, we demonstrated an important association between the CaOx deposition and graft loss, mainly due to chronic allograft nephropathy.

In summary, we can hypothesize that CaOx deposition may act as an additional non-immunologic factor in the pathogenesis of chronic allograft nephropathy, by injuring the tubular cell injury, and by eliciting interstitial fibrosis and loss of functioning nephron mass, in a setting of favorable conditions for its deposition. It is noteworthy to emphasize that the association of ATN and CaOx represent the worst combination for tubular cell injury and late graft recovery.

Acknowledgment

This study was supported by FAPESP—Fundação de Amparo à Pesquisa do Estado de São Paulo, CNPq—Conselho Nacional de Desenvolvimento Científico e Tecnológico, Fundação IMIPEN–Instituto Mineiro de Estudos e Pesquisas em Nefrologia, Fundação Oswaldo Ramos.

Ancillary