Factors Affecting Renal Allograft Function in Long-Term Recipients

Factors Affecting Renal Allograft Function in Long-Term Recipients

Factors Affecting Renal Allograft Function in Long-Term Recipients Peter C. Brazy, MD, John D. Pirsch, MD, and Folkert O. Belzer, MD • The natural his...

897KB Sizes 3 Downloads 57 Views

Factors Affecting Renal Allograft Function in Long-Term Recipients Peter C. Brazy, MD, John D. Pirsch, MD, and Folkert O. Belzer, MD • The natural history of renal allograft function in long-term recipients is not known. To characterize renal allograft function and the factors that affect it, we reviewed the records of all patients who received a renal allograft at the University of Wisconsin between 1965 and 1981 and selected those who had annual data on renal function for at least 10 years. We identified 155 patients-78 with living-related donors and 77 with cadaveric donors. All patients were adults receiving azathioprine and prednisone. Renal function was estimated by calculated creatinine clearances (Ccr), which correlated well with measured 24-hour creatinine clearances. The creatinine clearance data for each patient were plotted versus time. In 73% of patients, the creatinine clearance increased for several years before reaching a peak value. After the peak, the creatinine clearance declined in a linear manner. Stepwise regression analyses indicated that allografts from cadaveric donors had a greater increase in creatinine clearance from the value at year 1 to the peak than allografts from living-related donors (0.35 ± 0.25 v 0.21 ± 0.23 mL/s [21.4 ± 15.0 v 12.7 ± 13.8 mL/min); P < 0.001). The average time to reach the peak value of creatinine clearance was longer in cadaveric allografts (6.8 ± 3.5 v 4.6 ± 4.0 years; P < 0.001). Postpeak, the rate of decline in creatinine clearance was faster in cadaveric allografts than in living-related ones (0.06 ± 0.05 v 0.04 ± 0.04 mL/s/yr [3.50 ± 2.99 v 2.55 ± 2.16 mL/min/yr); P < 0.05). The presence of diastolic hypertension (average> 89 mm Hg) was associated with a more rapid rate of decline in creatinine clearance. The presence of diabetes mellitus or hypercholesterolemia was not associated with a more rapid decline in renal function. These data indicate that the natural history of renal allograft function in longterm recipients was to increase for several years and then to decline linearly. Allografts from cadaveric donors showed the most functional hypertrophy and had faster rates of decline. From these data, the average estimated life span of a successful renal allograft was 30 years for a cadaveric donor and 40 years for a living-related donor. © 1992 by the National Kidney Foundation, Inc. INDEX WORDS: Creatinine clearances; cadaveric donors; living-related donors; hypertension; diabetes mellitus; hypercholesterolemia.

T

HE MAXIMAL life span of a successful renal allograft is unknown. The major barriers to allograft survival are immune-mediated rejection,1 death of the patient with a functioning allograft from cardiovascular disease, infection or malignancy,2 and progressive renal failure from chronic rejection or intrinsic renal disease. This last barrier appears to cause late allograft loss at a rate of between 2% to 3% per year according to data from several transplantation centers. 2-4 The progressive loss of renal function in the long-term allograft may be mediated in part by the same nonimmune mechanisms that cause glomerulosclerosis and interstitial fibrosis in native kidneys.5-7 In native kidneys, systemic hypertension,8,9 glomerular hyperfiltration,1O and hypercholesterolemia 11 are associated with a more rapid rate of decline in renal function. The renal allograft is frequently exposed to systemic From the Departments 0/ Medicine and Surgery, University and the William S. Middleton Memorial Veterans Administration Medical Center, Madison, WI. Presented in part at the Annual Meeting o/the American Society o/Nephrology, Baltimore, MD, November 1991. Address reprint requests to Peter C. Brazy, MD, H4/51O, University 0/ Wisconsin Hospital and Clinics, 600 Highland Ave, Madison, W152792. © 1992 by the National Kidney Foundation, 1nc. 0272-6386/92/1906-0009$3.00/0

0/ Wisconsin,

558

hypertension 12 and hypercholesterolemia 13 and almost always has glomerular hyperfiltration. These and other factors may affect renal allograft function and determine its functional life span. Knowledge of the natural history of renal allograft function and identification of complicating factors may provide a key to prolonging allograft survival. The purpose of the present study was to characterize the natural history oflong-term renal allograft function and identify factors that affect its course. To this end, we examined the records of all renal transplant patients at the University of Wisconsin, Madison, who had annual data on renal allograft function for at least 10 years. We chose these criteria to exclude patients who suffered from subacute rejection and to have sufficient time to determine a rate of decline in renal function. Stepwise regression analyses of data from 155 patients indicated that the type of kidney donor and the presence of hypertension were significant factors in determining the time course of renal allograft function. METHODS For this study, we selected adolescent and adult patients (age range at time of transplant, 13 to 62 years) who had data on renal allograft function at least annually for 10 or more

American Journal of Kidney Diseases, Vol XIX, No 6 (June), 1992: pp 558-566

559

NATURAL HISTORY OF RENAL ALLOGRAFT FUNCTION

years. We reviewed the medical records of all patients who received a renal allograft at the University of Wisconsin, Madison, between January I, 1966 and December 31, 1980. In that time, 543 patients received 627 renal allografts, 430 from cadaveric donors and 197 from living-related donors. Eighty-four patients received an allograft for the second or third time. We identified 155 patients (25% of all patients) who had sufficient data to determine renal allograft function annually for 10 or more years. The outcome of unselected patients was allograft failure before 10 years (28% of all patients), death before 10 years (17%), and insufficient data to determine renal function annually (30%). From the medical record of the patient's annual visits to the University ofWisconsin Hospital and Clinics, we obtained general information about the patient (age, gender, type of transplant, medical problems, medications, age of donor kidney) and objective data (weight, blood pressure, serum creatinine, creatinine clearances, serum cholesterol). Data were collected through March I, 1991. The 155 selected patients included 93 men and 62 women. Seventy-eight patients had living-related donors, 77 patients had cadaveric donors, and 16 patients had received a prior allograft. All but one of the patients were white. The average age of the patient at the time of transplant was 32.5 years. The average age of the donor kidney was 28.8 years, with a range of I to 61 years. At the time of this study, the average duration of transplant function in these patients was 13.8 years, with a range of 10 to 24 years. Immunosuppression was achieved with azathioprine in all but three patients at an average dose of 100 mg/d and prednisone at an average dose of 11.5 mg/d. The cause of renal disease in the patients' native kidneys was chronic glomerulonephritis in most cases. Diabetes mellitus was present in 15% of the patients. Other frequent (5% to 10%) causes of native kidney failure were hypertension, polycystic kidney disease, and interstitial nephritis. We began our data collection as close as possible to the 1year anniversary of the transplant. This time was chosen because the prednisone dose was very near maintenance levels. In the second decade posttransplant, 15 of the selected patients developed end-stage renal disease and 10 died with a functioning allograft. These rates of late allograft and patient loss are comparable to reported ones. 2 Renal allograft function was evaluated by a calculated creatinine clearance (Ccr) rate. We used the Cockcroft and Gault equations l4 to calculate the creatinine clearance from the serum creatinine value, the patient's weight, age, and gender. We examined the validity of using these equations as an index of renal allograft function in a subset of our patients. In these 61 patients, Ccr data were compared with concurrent measurements of creatinine clearance with 24-hour urine collections on 208 occasions at intervals from I to 10 years posttransplant. A plot of the calculated Cockcroft-Gault creatinine clearance as a function of the measured 24-hour creatinine clearance is shown in Fig. I. The data are well described by linear regression (y = O.62 X + 26.0, r = 0.75) and are particularly good in the range of 0.33 to 1.33 mLls (20 to 80 mLI min). This result is similar to that observed in patients and normal subjects with two native kidneysl s and substantiates the use of these equations to estimate renal allograft function. Several medications such as cimetidine and trimethoprim cause an increase in serum creatinine values without altering

c:

1S

E

160

g

140

'e Ii

.. .

l!!

(3

.5 c:

i.., (.)

;

:; ~

lJ

0

o o

120

0

100



80

o

~o

60 40 2

o~o

00

20

40

0

0 0 00

0 00

0

~o ~o

0

0

00

60

80

100

120

140

160

180

Measured Creatinine Clearance; mil min

Fig 1. Relationship of calculated to measured values of creatinine clearance in renal allografts. Calculated Ccr data from a Cockcroft-Gault equation were compared with a simultaneously measured creatinine- clearance from a 24-hour urine collection on 208 occasions in 61 renal allograft patients. Data were collected at time points from 1 to 10 years posttransplant. Linear regression analysis of the data indicate a good correlation (y = 0.62x + 26.0, r = 0.75). To convert Ccr (mL/min) to Ccr (mL/ s), multiply by 0.01667.

the glomerular filtration rate. Only one of our patients was chronically on such a medication (cimetidine). Statistical analyses of the data on allograft function included (I) linear regression of data from individual patients to determine the rate of change of creatinine clearance before and after the peak value, (2) stepwise regression of data from the whole group of patients to determine which factors had a significant contribution to the variability of the outcome, and (3) comparisons between means of data from subgroups using an appropriate non parametric or parametric test. Statistical analyses were performed on a Macintosh computer (Apple Computer, Inc, Cupertino, CAl using Statview II software from Abacus Concepts, Berkeley, CA.

RESULTS

The ecr rates for each patient were plotted versus time. Inspection of these plots showed the presence of three patterns, which differed with respect to the time to peak ecr. In the first pattern (data from a typical patient shown in Fig 2A), the peak ecr occurred during the first 2 years posttransplant. Approximately 27% of patients followed this pattern. Their average peak ecr was 1.40 ± 0.25 (SD) mL/s (83.8 ± 14.9 mL/min) and the rate of decline in their ecr after the peak was linear and averaged 0.04 ± 0.03 mL/s/yr (2.1 ± 2.2 mL/min/yr). In the second pattern (data from a typical patient shown in Fig 2B), the ecr increased for several years after transplant, reached a peak value, and then declined in a linear manner. The majority of patients (63%) fol-

BRAZY, PIRSCH, AND BELZER

560

lowed this pattern. Their Ccr at 1 year averaged 1.23 ± 0.37 mL/s (73.8 ± 22.0 mL/min). Their Ccr increased slowly to a peak value of 1.57 ± 0.45 mL/s (93.9 ± 26.7 mL/min) at an average

80

A

y _ 78.147 - 1.9469x RA2 - 0.915

70



c

E E

s

·•

60

c

I!

l!

u

50

.5 .5



! u

40

30

0

10

30

20

Time, yrs

80

B

••

• • • • .a. a " •

70 c

E E

s c

I!

:



60



"a

II

u



.5 c

ii !

u

-. •

50

40

10

0

30

20 Time, yrs

110 y - 75.565 + 2.6506x

RA2 - 0.818

time of 6.5 ± 2.8 years posttransplant. In these patients, the rate of decline in Ccr postpeak averaged 0.06 ± 0.05 mL/s/yr (3.5 ± 3.0 mL/minl yr). In the third pattern (data from a typical patient shown in Fig 2C), the Ccr increased during the entire time of observation. These patients represented 10% of the total population. As a group, their Ccr at 1 year averaged 1.17 ± 0.40 mL/s (70.1 ± 24.0 mL/min), their average rate of increase in Ccr was 0.04 ± 0.02 mL/s/yr (2.1 ± 1.1 mL/min/yr), and their last Ccr averaged 1.73 ± 0.46 mL/s (103.6 ± 27.4 mL/min). Presumably, these patients will eventually display the same Ccr pattern as the second group of patients. Because the patient's weight was a factor in the calculation of Ccr, an increase i~ weight would contribute to the increase in Ccr. We determined the average weight of patients at the time of the peak Ccr and found it to be 107.1 % ± 10.0% of the weight at the time of the initial Ccr. This weight gain would contribute a modest part of the 27% to 48% increases in Ccr that were observed . Rather than making comparisons between these groups, we examined all the data with stepwise regression analyses to determine whether patients' gender, patients' age, donors' age, or type of kidney (living-related or cadaveric donor) would explain some of the variability of the course of renal function in long-term renal allografts. The patient's gender and age were significant determinants of the variability of the Ccr at 1 year posttransplant. The mean value of Ccr at 1 year for women (1.17 ± 0.36 mL/s [70.4 ± 21.7 mL/min]) was significantly lower than the mean value for men (1.33 ± 0.33 mL/s [80.0 ± 19.7 mL/min]; P< 0.02). When the data were sorted by the patients' age at the time of transplant, the mean Ccr at 1 year for patients in the

100

~

OIl'



E

ac

90

·

80

j

u

c "E

]



70 0

2

4

6 Time, yrs

8

10

12

Fig 2. (A) Plot shows annual data of calculated Ccr in a typical renal allograft patient whose peak Ccr occurred in the first 2 years posttransplant. Linear reg ression analysis of the data points indicate a linear rate of decline (y = -1.9x + 78.1, r = 0.95). (8) Plot shows annual data of calculated Ccr in a typical patient whose peak Ccr occurred several years posttransplant. (C) Plot shows annual data of calculated Ccr in a typical patient whose peak Ccr had not yet occurred. Linear regression analysis of the data points indicate a linear rate of increase (y = 2.6x + 75.6, r = 0.9). To convert Ccr (mL/ min) to Ccr (mL/s), multiply by 0.01667.

561

NATURAL HISTORY OF RENAL ALLOGRAFT FUNCTION

youngest quartile (age range, 13 to 23 years) was significantly greater (1.30 ± 0.33 mLls [78.2 ± 19.7 mL/min]) than the mean value (U5 ± 0.31 mLls [69.1 ± 18.6 mL/min]; P < 0.05) of patients who were in the oldest quartile (age range, 40 to 62 years). These observations are entirely consistent with the use of the Cockcroft and Gault equations in which the Ccr is reduced by 15% in women and by 7% for each decade of age. 14 A majority of our patients (68%) achieved a peak value of Ccr that was in the "normal" range for people with two kidneys (1.33 to 2.17 mLls [80 to 130 mL/min] for men, 1.25 to 2.00 mLI s [75 to 120 mL/min] for women, values from University of Wisconsin Hospital and Clinics Laboratory). One fourth of the patients had a peak Ccr that was less than the normal range, and 6% of the patients had a peak Ccr that was above normal. Stepwise regression analysis of the peak Ccr value identified the age of the donor kidney and the Ccr of the patient at I year as being significant variables. We sorted the peak Ccr data into quartiles by the age of the donor kidney at the time of transplant (Fig 3) and applied an analysis of variance (ANOVA) and a Fisher's protective least-significant difference test to compare mean values between quartiles. The analyses indicated that the mean value of peak Ccr for patients with a donor kidney from the youngest quartile, age range 1 to 20 years, was significantly greater (1.70 ± 0.47 mLls [102.0 ± 28.0 mL/min]) than that for patients with a donor kidney from the third quartile, age range 27 to 37 years (peak Ccr, 1.40 ± 0.37 mLls [84.2 ± 22.2 mL/min]; P < 0.05). The mean value for peak Ccr of the oldest quartile (1.53 ± 0.33 mLI s [91.9 ± 19.5 mL/min]) was not significantly different from that of the youngest (P = 0.08). In a similar manner, we examined the effect of the Ccr at I year on the peak Ccr. Patients with the lowest Ccr at I year, range 0.30 to 1.05 mLls (18 to 63 mL/min), had an average peak Ccr of U8 ± 0.42 mLls (70.8 ± 15.1), which was significantly lower than the peak Ccr (1.97 ± 0.38 mLI s [118.0 ± 22.5 mL/min]; P < 0.05) for patients with the highest Ccr at year I, range from 1.48 to 2.45 mLls (89 to 147 mL/min). Thus, the peak value of Ccr was inversely related to the age of the allograft donor and directly related to the Ccr I year posttransplant.

110

'00 c

~

E 90

3

I! eo t 1 (J

70

60

L - _...... , -.20...&._......21-...L -......- -.... ' 1-......_,..•..1,1 - . 27 3

IIJE OF DONOR KlJNE.Y. YEARS

Fig 3. Effect of kidney age on peak Ccr. Mean ± SO values lor peak Ccr are shown lor each quartile of population sorted by the age of the donor kidney at time of transplant. ANOVA and Fisher's least-significant difference tests indicated that group 3 (age range, 27 to 37 years) is significantly less (P < 0.05) than group 1 (age range, 1 to 20 years). To convert Ccr (mL/min) to Ccr (mL/s), multiply by 0.01667.

Next we examined the data to determine which factors might contribute to the increase in Ccr (the difference between the peak and the I-year value of Ccr). A stepwise regression identified only the type of kidney donor as being significant. The average increase in Ccr was 0.21 ± 0.23 mLI s (12.7 ± 13.8 mL/min) for patients receiving a kidney from a living-related donor and 0.36 ± 0.25 mLls (21.4 ± 15.0 mL/min) for patients receiving a kidney from a cadaveric donor (P < 0.0002 by unpaired t test). Stepwise regression analysis of factors affecting the variability in the time to reach the peak Ccr indicated the type of kidney donor and the Ccr at I year as being significant. Patients with a cadaveric kidney had an average time of 6.8 ± 3.5 years to reach the peak Ccr, which was significantly greater than the time for patients with a living-related donor (4.6 ± 4.0 years, P < 0.0003 by unpaired t test). Patients in the lowest quartile for Ccr at I year had an average time of 7.2 ± 3.4 years to reach the peak Ccr, as compared with an average of 4.9 ± 3.5 years for patients in the highest quartile for Ccr at I year (P < 0.05 by an ANOVA and a Fisher's protective least-significant difference test). After reaching a peak Ccr, the Ccr decreased in a linear manner. We examined the data to determine which factors might contribute to the

BRAZY, PIRSGH, AND BELZER

562 40

0

• i

...

i

'0

Living-Related Kidneys



30

Cedavarlc Kidney•

20

j

E

~

10

01.0

1.1 2.0

2.13.0

3.1 4.0

4.15.0

6.1-

5.16.0

7.0

7.18.0

>8.0

Fig 4. Frequency distribution of postpeak rate of decline In Ccr. The frequency of different rates of decline in Ccr are shown for patients receiving either a living-related kidney (0) or a cadaveric kidney (~). To convert Ccr (mL/ min) to Ccr (mL/s), multiply by

0.01667_

Rate 01 Doctlne In Ccr; mil min I y

.

min/yr); P < 0.05 by Mann-Whitney Utest). In a separate stepwise regression analysis, we examined the relationship of the Ccr at 1 year, the peak Ccr, and the time to peak to the rate of decline in Ccr. This analysis indicated that none of these variables had a significant effect in determining the variability of the rate of decline in Ccr. The characteristics of renal allograft function and patients with either a living-related or a cadaveric donor are listed in Table 1. Patients with cadaveric kidneys differed from those with livingrelated kidneys in the time to peak (longer), the rate of decline in Ccr (faster), the age of the donor (younger), their age at the time of transplant

variability in the rate of decline. Stepwise regression analysis of patients' gender, patients' age, kidney age, and type of donor indicated that only the type of kidney had a significant effect on the rate of decline. A frequency distribution plot of the rate of decline in Ccr for both living-related and cadaveric kidneys showed a skew distribution of both sets of data (Fig 4). Therefore, nonparametric tests were used to compare the mean values for rates of decline in Ccr. Such an analysis indicated that the mean rate of Ccr decline in patients with cadaveric kidneys (0.06 ± 0.05 mL/ s/yr [3.60 ± 2.98 mL/min/yr]) was significantly greater than that in patients with living-related donors (0.04 ± 0.03 mL/s/yr [2.35 ± 1.98 mL/

Table 1. Characteristics of Renal Allograft Function by Type of Donor Type of Donor Uving-Related

No. of patients Initial Gcr, mL/s (mL/min) Peak Gcr, mL/s (mL/min) Time to peak, yr Rate of Gcr decline, mL/s/yr (mL/min/yr) Age of donor, yr Age of recipient, yr Maintenance prednisone dose, mg/d Systolic BP, mmHg Diastolic BP, mmHg

1.32 1.53 4.6 0.04 35.0 29.6 10.1 129.6 82.9

± ± ± ± ± ± ± ± ±

78 0.33 (79.3 ± 19.5) 0.36 (92.0 ± 21.5) 4.0 0.04 (2.55 ± 2.16) 10.9 11 .1 2.4 12.6 8.8

Cadaveric

77 1.22 1.58 6.8 0.06 22.5 35.4 12.3 131.4 84.1

± ± ± ± ± ± ± ± ±

0.37 (73.0 ± 22.1) 0.45 (94.4 ± 26.8) 3.5" 0.05" (3.50 ± 2.99) 9.3" 12.3" 2.9" 11.5 6.6

NOTE. Values are means ± SO. " Indicates values of cadaveric group that are significantly different (P < 0.05) from values of living-related group by Student's unpaired t test for parametric data or Mann-Whitney U test for non parametric data.

563

NATURAL HISTORY OF RENAL ALLOGRAFT FUNCTION

(older), and the latest dose of prednisone (higher). They did not differ with regard to blood pressure measurements in clinic. We used these parameters of renal allograft function to calculate the average expected functional "life span" for cadaveric and living-related renal allografts. If we assume that renal allografts have failed at a Ccr of 0.17 mL/s (10 mL/min), then the average life span for an allograft would be 30 years for a cadaveric donor and 40 years for a living-related donor. In chronic progressive renal disease, systemic hypertension, hyperglycemia, and hypercholesterolemia have individually been associated with more rapid rates of decline in renal function. We examined the effects of these conditions on the rate of decline in Ccr in our population of renal allografts. The average value of diastolic blood pressure, measured in the renal transplant clinic, was determined for each patient and the data were sorted into quartiles on the basis of this value (Fig 5). ANOVA indicted a significant difference among the quartiles. Nonparametric comparisons with a Wilcoxon signed-rank test indicated that patients in the highest quartile, whose mean diastolic blood pressure ranged from 89 to 98.8 mm Hg, had an average rate of decline in Ccr of 0.07 ± 0.06 mL/s/yr (4.00 ± 3.28 mL/min/yr), which was significantly greater (P < 0.05) than that of the other three quartiles. The number of antihypertensive medications that a patient was taking may be another indicator of the severity of hypertension. The usual treatment pattern in our hypertensive patients was to start with a diuretic and add hydralazine and a /3-blocker as needed. However, there was only a nonsignificant tendency for the patients on two or more medications to have a faster rate of decline (0.06 ± 0.05 mL/s/yr [3.48 ± 2.94 mL/min/yr); N = 72) than the patients on one or no medications (0.05 ± 0.05 mL/s/yr [2.76 ± 2.81 mL/min/yr); N = 83; P = 0.08 by Mann-Whitney Utest). There were 23 patients with diabetes mellitus (10 with type I and 13 with type II). The average rate of decline in Ccr in these patients (0.06 ± 0.05 mL/s/yr [3.69 ± 3.25 mL/min/yr)) was not significantly different from that of nondiabetic patients (0.05 ± 0.04 mL/s/yr [2.88 ± 2.48 mL/min/yr). The nonfasting serum cholesterol was measured as part of a standard panel of clinical chemistry tests. Based on the average value

:a..

iE

1.0

i

I ...

§

I ..~ !

J

3 .•

•.•

60-78

70-84

85·8e

.....

MEANDISTOLIC

BLOOD PRESSURE mmHg

Fig 5. Effect of diastolic blood pressure on rate of decline in Ccr. Mean ± SO values for the postpeak rate of decline in Ccr are shown for each quartile of the population sorted by the mean value of diastolic blood pressure. ANOVA and Wilcoxon signed-rank test indicated that the mean of group 4 (blood pressure range, 89 to 99 mm Hg) was significantly faster than that of the other three quartiles (P < 0.05). To convert Ccr (mL/min) to Ccr (mL/s), multiply by 0.01667.

of serum cholesterol, patients were divided into three categories: high, cholesterol greater than 6.2 mmol/L (240 mg/dL); borderline, cholesterol between 5.2 and 6.2 mmol/L (200 and 240 mg! dL); and normal, cholesterol less than 5.2 mmol/ L (200 mg/dL). The average rate of decline in Ccr for patients with high cholesterol was 0.05 ± 0.05 mL/s/yr (2.94 ± 2.76 mL/min/yr) (N = 53), which was not significantly different from the average value for patients with normal cholesterol (0.04 ± 0.03 mL/s/yr [2.22 ± 1.86 mL/ min/yr); N = 24). Thus, neither the presence of diabetes mellitus nor a high serum cholesterol was associated with a more rapid rate of decline in Ccr. DISCUSSION

The basic question that we wished to address is what happens to renal allograft function in the absence of acute or subacute rejection or early death of the patient. A recent retrospective study of such patients from the University of Minnesota reported that the expected half-life for allograft survival was 19.4 years for living-related donors and 17.5 years for cadaveric donors. 2 Our data estimated the average allograft survival rate as being 40 years for living-related donors and 30 years for cadaveric donors. The two estimates are essentially the same

564

and consistent with the 2% to 3% annual rate of allograft loss reported by others. 3,4 Less information is available regarding what happens to renal allograft function before failure. Our data address this question. We observed the following: (1) The Ccr of most renal allografts increased for several years posttransplant. (2) The Ccr peaked at a value that was in the range of normal renal function. (3) Following this peak, the Ccr of renal allografts declined linearly at a rate that averaged 0.04 mL/s/yr (2.4 mL/min/yr) for patients with living-related kidneys and 0.06 mL/s/yr (3.5 mL/min/yr) for patients with cadaveric kidneys. Additionally, we identified factors that determined some of the variability in these changes in renal allograft function. A novel observation in our study was the fact that Ccr in most renal allografts continued to increase for several years. This result was not unexpected. A prior study from the University of Wisconsin reported serial measurements of inulin or iothalamate clearances in a group of kidney donors up to 4 years following nephrectomy.16 This prospective study showed that glomerular filtration rates increased continuously during that time and that the amount of increase was greatest in the youngest donors. A retrospective study from the Mayo Clinic reported serial measurements of serum creatinine in kidney donors 10 to 20 years postnephrectomy.17 These data showed a significant decrease in serum creatinine with time, consistent with a prolonged process of increasing glomerular filtration rate. In renal allograft recipients, the increase in creatinine clearance should be associated with an increase in glomerular size. Renal biopsies were rarely performed in our study patients, so we did not have sufficient tissue for this type of analysis. The University of Minnesota grouplS has examined glomerular histopathology in biopsy specimens from a series of renal allografts and found significant hypertrophy of glomerular tissue in both rejecting and nonrejecting allografts. They reported that the glomerular surface area, a determinant of glomerular filtration rate, increased as a function of time after transplant. IS These anatomic observations support the hypothesis that the increase in creatinine clearances is due to glomerular hypertrophy and that the process may occur over several years in the renal allograft. In most patients, the Ccr increased until it had reached a value that was approximately normal for

BRAZY, PIRSCH, AND BELZER

people with two kidneys. Factors that had a positive correlation with the peak Ccr value were the age of the donor kidney and the value of Ccr at 1 year posttransplant. It is known that the capacity for functional hypertrophy in kidney donors declines with age. 16,19 In our study, allografts from young donors were more likely to have a higher peak Ccr than allografts from older donors, without regard for the age of the recipient. This finding suggests that the potential for functional hypertrophy resides more in the allograft than in the environment of the host. The value of Ccr at 1 year directly correlated with the peak Ccr. The Ccr at 1 year probably reflected the number of functional nephrons in the allograft, a factor that would correlate directly with potential peak value ofCcr. Our analyses also indicated that the type of kidney and the Ccr at 1 year were predictors of the duration of time to reach the peak Ccr. Patients with cadaveric renal allografts had a greater increase in Ccr from the value at year 1 posttransplant to the peak Ccr and a longer duration of increasing Ccr than patients with livingrelated renal allografts. Allografts with a low Ccr at 1 year also had a longer duration before reaching their peak Ccr. An explanation for this association may be that these allografts have a reduced number of functional nephrons as a result of either subacute rejection or preservation injury. It is unlikely that the method of allograft preservation contributed significantly to the variability in cadaveric allograft function after the first year. A large, prospective, clinical study found that neither the method nor duration of renal allograft preservation had an effect on allograft survival or function at 1 year.20 Our data indicate that unknown factors stimulate functional renal hypertrophy for many years after transplantation. These growth factors may be the same ones that are present in native kidneys which have lost nephrons to inj1ll'1,21,22 or may be released as a part of the chronic rejection process. 7 Our data indicated that the rate ofdecline ofCcr in long-term renal allografts was quite variable (Fig 4). Twenty percent of the patients had rates of decline that were less than 0.02 mL/s/yr (1.0 mL/ min/yr). Forty-eight percent of the patients had rates of decline that were greater than 0.03 mL/S/ yr (2.0 mL/min/yr), a rate that is twice that of normal men. 23 The frequency data in Fig 4 did not follow a normal distribution. Therefore, it is likely that multiple factors have a significant role in determining the rate of decline. Is the accelerated rate

565

NATURAL HISTORY OF RENAL ALLOGRAFT FUNCTION

of decline in allograft function due to the hemodynamic consequences of hypertrophy in a single kidney, to the immune-mediated injury of chronic rejection, or both? The rate of decline of renal function in patients with normal single kidneys is not well documented. Follow-up studies of transplant donors indicated that renal function was preserved,19,23,24 but did not provide enough data to determine the duration of increasing renal function l6,17 or the subsequent rate of decline in creatinine clearance. In our study, the level of Ccr at year 1 did not predict the rate of decline in Ccr. Patients with low values did not show faster rates ofdecline in Ccr. Similarly, the peak level ofCcr did not predict the rate of decline in Ccr. Patients with a marked degree of functional hypertrophy did not show a faster rate of decline in Ccr. Thus, the early level of renal allograft function does not predict the subsequent rate of declining function. Our main finding was that cadaveric allografts had significantly faster rates of decline in function as compared with living-related allografts (Table 1). Presumably, the main difference between these two groups was the higher incidence of chronic rejection in the cadaveric allografts. In a recent study from the University of Minnesota, the renal function of 22 patients with pathologically proven chronic allograft rejection declined at an average rate of 0.08 mL/s/yr (4.8 mL/min/yr).25 This rate of decline (>0.07 mL/s/yr [4.0 mL/min/yr]) was seen in 25% of our patients (Fig 4). Although we did not have tissue to document the presence of chronic rejection, this diagnosis is the most likely explanation for those patients with rapidly declining function. Systemic hypertension has clearly been associated with an accelerated rate of progressive renal failure in patients with chronic renal disease6,9.26 and with diabetic nephropathy.8,27 Control ofsystemic hypertension slowed the rate of progressive renal failure in many cases. The present studyexamined the relationship between systemic hypertension and the rate of decline in Ccr in long-term renal allografts. Hypertension was not a common problem in this population, only 25% of patients exhibited persistent diastolic hypertension. How-

ever, in those patients, the rate of decline in Ccr was 50% greater than in patients with controlled hypertension (Fig 5). These data do not indicate whether hypertension was the cause or a result of a more rapid decline in Ccr because hypertension is frequently present in renal allograft rejection and may appear as a consequence of rapidly declining renal function. In this regard, the recent study by Modena et al 25 reported that in patients with chronic rejection, faster rates of decline in renal function were found in patients with higher blood pressures. Furthermore, a diastolic blood pressure greater than 90 mm Hg in individual patients was associated with a faster rate of decline. 25 Thus, the effects of hypertension on progressive renal disease in renal allografts and in diseased native kidneys appears to be similar. . The present study characterized the natural history of allograft renal function for kidneys that have functioned for more than 10 years. This study indicated the utility of creatinine clearance versus time plots to identify trends in allograft function. Such plots may document therapeutic benefits or adverse effects of specific risk factors, ie, hypertension. The results of this study were limited by the fact that the patient population was homogeneous, young, white Americans and that the immunosuppressive therapy was azathioprine and prednisone. The introduction of cyclosporine and OKT3 has markedly improved rates of the allograft survival during the first 5 years posttransplant. 28 The effects on long-term renal allograft function of improved immunosuppression of early rejection episodes and of chronic cyclosporine nephrotoxicity29 remain to be determined. When such data are available, the type of analysis used in this study may serve as a benchmark to evaluate the overall effects of new therapies on renal allograft function. ACKNOWLEDGMENT The authors thank the University of Wisconsin Renal Transplant Coordinators and Nurse Oinicians, particularly Jean Esch, for their diligent efforts in patient care and record keeping. We thank Drs Aaron Friedman and Richard Rieselbach for their helpful comments in reviewing the manuscript.

REFERENCES I. Mahony JF: Long-term results and complications of transplantation: The kidney. Transplant Proc 21:1433-1434, 1989

2. Fischel RJ, Payne WD, Gillingham KJ, et al: Long-term outlook for renal transplant recipients with one-year function. Transplantation 51: 118-122, 1991

566 3. Kirkman RL, Strom TB, Weir MR, et al: Late mortality and morbidity in recipients of long-term renal allografts. Transplantation 34:347-351, 1982 4. Braun WE: Long-term complications of renal transplantation. Kidney Int 37:1363-1378, 1990 5. Feehally J, Bennett S, Harris K, et al: Is chronic renal transplant rejection a nonimmunological phenomenon? Lancet 2:486-488, 1986 6. Klahr S, Schreiner G, Ichikawa I: The progression of renal disease. N Engl J Med 318:1657-1666, 1988 7. Fellstrom B, Larsson E, Tufveeson G: Strategies in chronic rejection of transplanted organs: A current view on pathogenesis, diagnosis, and treatment. Transplant Proc 21: 1435-1439, 1989 8. Mogensen CE: Long-term anti-hypertensive treatment inhibiting progression of diabetic nephropathy. Br Med J 285: 685-688, 1982 9. Brazy PC, Stead WW, Fitzwilliam JF: Progression of renal insufficiency: Role of blood pressure. Kidney Int 35: 670-674, 1989 10. Brenner BM: Nephron adaptation to renal injury or ablation. Am J PhysioI249:F324-F337, 1985 II. Kasiske BL, O'Donnell MP, Schmitz PG, et al: Renal injury of diet-induced hypercholesterolemia in rats. Kidney Int 37:880-891, 1990 12. Cheigh JS, Wang J, Fine P, et al: Hypertension and decreased graft survival in long-term kidney transplant recipients. Transplant Proc 17:174-175, 1985 13. Kasiske B, Umen AJ: Persistent hyperlipidemia in renal transplant patients. Medicine (Baltimore) 66:309-316, 1987 14. Cockcroft D, Gault M: Prediction of creatinine clearance from serum creatinine. Nephron 16:31-41, 1976 15. Lemann Jr J, Bidani A, Bain RP, et al: Use of the serum creatinine to estimate glomerular filtration rate in health and early diabetic nephropathy. Am J Kidney Dis 16:236243,1990 16. Bonner G, Shelp WD, Newton M, et al: Factors influencing the increase in glomerular filtration rate in the re-

BRAZY, PIRSCH, AND BELZER

maining kidney of transplant donors. Am J Med 55:169-174, 1973 17. Anderson C, Velosa J, Frohnert P, et al: The risks of unilateral nephrectomy: Status of kidney donors 10 to 20 years postoperatively. Mayo Clin Proc 60:367-374, 1985 18. Kasiske B, Kalil R, Lee H, et al: Histopathologic findings associated with chronic, progressive decline in renal allograft function. Kidney Int 40:514-524, 1991 19. Slack T, Wilson D: Normal renal function. Cn and Cpah in healthy donors before and after nephrectomy. Mayo Clin Proc 51 :296-300, 1976 20. Spees EF, Vaughn WK, Mendez-Picon G, et al: Preservation methods do not affect cadaver renal allograft outcome. The SEOPF prospective study 1977-1982. Transplant Proc 16:177-179,1984 21. Hakim RM, Lazarus JM: Progression of chronic renal failure. Am J Kidney Dis 14:396-401, 1989 22. Jacobson HR: Chronic renal failure: Pathophysiology. Lancet 338:419-423, 1991 23. Lindeman RD, Tobin JD, Shock NW: Association between blood pressure and the rate of decline in ~enal function with age. Kidney Int 26:861-868, 1984 24. Fontino S: The solitary kidney: A model of chronic hyperfiltration in humans. Am J Kidney Dis 13:88-98, 1989 25. Modena FM, Hostetter TH, Salahudeen AK, et al: Progression of kidney disease in chronic renal transplant rejection. Transplantation 52:239-244, 1991 26. Brazy PC, Fitzwilliam JF: Progressive renal disease: role of race and antihypertensive medications. Kidney Int 37: 1113-1119, 1990 27. Parving HH, Smidt UM, Andersen AR, et al: Early aggressive antihypertensive treatment reduces rate of decline in kidney function in diabetic nephropathy. Lancet I: 117 51177, 1983 28. Stratta RJ, D'Alessandro AM, Hoffmann RM, et al: Cadaveric renal transplantation in the cyclosporine and OKT3 eras. Surgery 104:606-615, 1988 29. Myers BD, Sibley R, Newton L, et al: The long-term course of cyclosporine-associated chronic nephropathy. Kidney Int 33:590-600, 1988