Protocol Core Needle Biopsy and Histologic Chronic Allograft Damage Index (CADI) as Surrogate End Point for Long-Term Graft Survival in Multicenter Studies

Materials and Methods

Statistical Analyses

Spearman correlation coefficient was used to evaluate the correlation between the two pathologists in quantitation of CADI (Figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Correlation between the two pathologists in quantitation of the Chronic Allograft Damage Index (CADI) in 621 representative protocol core needle biopsies. Spearman correlation coefficient 0.85.

Multivariate logistic regression analysis was employed to identify independent predictors of 3-yr graft loss (including patient death). The model included variables that were significant in a univarate analysis (P < 0.05) or were of biologic interest. Independent variables included in the model are listed in Table 2. Odds ratios (OR) and corresponding 95% confidence intervals were calculated. P-values less than 0.05 were considered to indicate statistical significance.

View this table:

• View inline
• View popup

Table 2. Parameters correlating (P) with 3-yr graft loss or patient death in the combined study^a

Multivariate linear regression analysis was performed to identify independent predictors of the 1-yr CADI score (Table 3). P values less than 0.05 were considered to indicate statistical significance.

View this table:

• View inline
• View popup

Table 3. Parameters correlating (P) with high CADI score at 1 yr in the US, Tricontinental, and combined studies^a

A χ² test for multiple proportions was conducted for categorical data (Table 4).

View this table:

• View inline
• View popup

Table 4. Correlation between 1-yr CADI score and patient/graft survival at 3 yr (patient with representative protocol biopsy)

Results

Progression of the CADI Score During Three Years Posttransplantation

Whereas the mean serum creatinine in all patients at 1 yr was still close to normal, 1.6 ± 0.7 mg/dl (mean ± SD), the mean CADI score had already progressed to 3.3 ± 1.8 (Figure 2). At 1 yr, only 20% of patients had elevated (>l.5 mg/dl) serum creatinine, whereas 60% of the 1-yr biopsies demonstrated an elevated, (>2.0) (3) CADI score.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. The histogram showing the distribution of CADI score and serum creatinine (mg/dl) at 1 yr, and the number of patients who fell within each interval. Elevated values are shaded.

The progression of the CADI score during the first 3 yr is shown in Table 1. In the total cohort of 621 representative protocol biopsies, the mean CADI score was 1.25 ± 1.13 (SD) at baseline (111 biopsies), 3.27 ± 1.82 at 1 yr (302 biopsies), and 4.10 ± 2.15 (208 biopsies) at 3 yr. We also determined separately the progression of the CADI score in those patients within the cohort who were biopsied two or three times during the study (Table 1). The baseline and progression of the CADI score in these subgroups was not any different of the total cohort. Thus the histologic changes relevant to chronic rejection developed fastest during the first year, approximately 2 CADI units/yr, but less rapidly during the subsequent 2 yr of follow up, 0.5 CADI units per year.

View this table:

• View inline
• View popup

Table 1. Summary of baseline, 1-yr, and 3-yr Chronic Allograft Damage Index (CADI) scores

Discussion

In a recent analysis of UNOS Registry data from 1988 to 1996, 1-yr cadaveric renal allograft survival has improved from 75.7 to 87.7%, as has transplant half-life after the first year from 7.9 to 13.8 yr (18). On the other hand, the half-life of transplants with one or more acute episodes of rejection improved only modestly, from 9.1 to 11.1 yr (18). Here, acute vascular rejection (19), often manifesting as steroid-resistant rejection with a strong humoral component (20), is a major negative prognostic indicator. With the decrease of graft loss during the first year, late graft loss after the first year has become the major reason of renal allograft failure.

The most frequent reason for late cadaveric renal allograft loss is death with a functioning transplant (53%) followed by chronic allograft rejection (36%) (21). The main cause of death in patients with functioning transplant is cardiovascular (22). Thus the two main reasons for late allograft failure, chronic rejection, manifesting as fibroproliferative intragraft vascular disease (23), and accelerated atherosclerosis, are closely related and may potentially respond to similar therapy.

There is ample evidence from eight single-center studies (3–10) that incipient histologic changes characteristic of chronic rejection are visualized in the transplant before the transplant function deteriorates. Four of these studies have employed Banff histologic criteria in their evaluation (6–8,10), one study the amount of collagen/fibrosis in the biopsy (9), and the remaining three a sum score of various findings in the graft interstitium, glomeruli, vessels, and tubuli (3–5). The CADI score is the most extensive of these in regard to the number of parameters quantitated and, in contrast to the Banff criteria for chronic rejection, takes also into account the vascular and glomerular changes in the biopsy.

Implementation of protocol core needle biopsy in a multicenter study was met with two difficulties. There was clinical concern at performing protocol biopsies in a well-functioning transplant. Thus, only about one third of the centers complied with this, though within these centers all of the participating patients were biopsied at 1 yr (unless there was a clinical contraindication to performing the biopsy). Second, to monitor the progression of the lesion, two or preferably three biopsies would be needed from a single patient, but only five centers complied with two biopsies and only two centers were willing to biopsy three times. Thus such follow-ups are clearly underrepresented in this study. The biopsies were mostly of good quality; 86% contained at least seven glomeruli, and the timing of the biopsy procedures followed closely the protocol.

Both trials, (US and Tricontinental), were originally planned for 12 mo of follow-up, but they were extended thereafter to 3 yr. The original sample sizes chosen were adequate to detect a reduction in acute rejection rate at 6 mo. The studies were not designed to test the hypothesis of whether chronic rejection or graft loss at 3 yr were impacted; therefore, it is not surprising that there were insufficient numbers of patients in these trials to address these factors. For this reason and because the patients who provided biopsy data no longer represented the total population, the effect of treatment (i.e., who received MMF versus placebo/azathioprine) was not addressed. The similar design of the two studies, however, enabled us to pool the data, when investigating the predictive value of protocol biopsy, as scored by the CADI criteria, for graft outcome, keeping in mind that the pool of patients with available biopsy data does not necessarily represent the total population in the two studies.

Most importantly, this study demonstrates that it is possible to perform a multicenter protocol biopsy-based trial with centralized biopsy scoring if the participating centers are selected for their willingness to perform the protocol biopsies. The study also confirmed the predictive value of the biopsy to subsequent graft outcome in a multicenter trial. Importantly, no major complications were observed in this study and no fatalities were reported as consequence of biopsy.

The results also show that high 1-yr CADI score correlates with known risk factors to chronic rejection, including donor age (24–26), acute rejection during first year (27,28), and CMV infection (29) (only in the separate studies but not in combined study). The fact that other frequently cited risk factors, such as cold ischemia time, delayed graft function, number of HLA mismatches, and others (for references, see references 30 and 31), did not correlate may be explained by too short interval between the transplant and the biopsy.

Additionally an elevated CADI score at 1 yr correlates with graft survival at 3 yr, both in patients with normal and elevated serum creatinine at the time of the biopsy. In other words, histopathologic findings in protocol biopsy are more sensitive markers than change in transplant function, as detected by serum creatinine level. At 1 yr, only 20% of patients had elevated (>l.5 mg/dl) serum creatinine, whereas 60% of the 1-yr biopsies demonstrated already an elevated (i.e. >2.0) (3) CADI score. When the clinical and biopsy status at 1 yr was correlated with the clinical outcome at 3 yr, not a single patient lost their graft in the low CADI group with normal serum creatinine, six patients (4.6%) lost their graft in the in the elevated CADI group also with normal serum creatinine, and (not unexpectedly) 17 patients (16.7%) lost their graft in the high CADI group displaying already an elevated serum creatinine level (P < 0.001). When death with a functioning graft was censored, the outcome was no lost grafts in the low CADI group, four lost grafts in the elevated CADI group, and 12 lost grafts in the high CADI group (P < 0.001). Although this analysis is based on a small cohort of patients with evaluable biopsy data at 1 yr, that the number of lost grafts at 3 yr was still small, and that the results need confirmation after 5 to 10 yr follow-up, the trend is quite clear. Our findings in the present study and those reported earlier, suggest that protocol core needle biopsy and CADI can be used as surrogate end point in primary prevention trials and possibly to identify the cohort of patients at risk for chronic rejection in secondary intervention trials, unless the 1-yr time point is too late to interfere with the disease.

We observed a rapid progression of the CADI from baseline to 1 yr. The pace slowed down during the subsequent years. The fast progression of CADI during the first year is compatible with the paradigm that chronic rejection represents graft response to cumulative injury, occurring mostly during the first year, regardless whether it is immunologic or non-immunologic in origin (23).

The baseline CADI in this study indicates that most of the transplants were satisfactory in quality, with a mean CADI at baseline of only 1.25. The range (0 to 5.3) shows, however, that some suboptimal transplants were also included in the study. This observation emphasizes the need to obtain a baseline biopsy to calculate the progression of the histologic manifestations during the first year. Assessment of the ΔCADI, rather than an isolated measure of the CADI at 1 yr, is necessary to accurately differentiate the outcome between patients destined to progress (progressors) and those unlikely to do so (non-progressors). The number of grafts lost at 3 yr was too small to make meaningful comparisons between progressors and non-progressors.

The histologic findings characteristic of acute rejection and histologic sequelae of an immunologic insult do not appear to be different in the MMF era (where the data in this manuscript is derived from) compared with those resulting from the use of newer agents that have resulted in lower rejection rates. Therefore, we believe that the conclusion that higher CADI score leads to poorer graft survival is still valid.

Taken together, our results in two multicenter trials confirm the results of several single-center studies that early histologic alterations, quantitated here as CADI, predict chronic rejection, even when the graft function is still normal. This observation will make it possible to use the approach not only as a surrogate end point in prevention trials, but also to identify the patient cohort at risk for intervention trials. One important lesson of this study, however, is that it is more meaningful to investigate the progression of the lesion (ΔCADI) rather than the intensity of alterations at a single given time point.