- Email: [email protected]
Effect of Pregnancy on Long-Term Function of Renal Allografts
Effect of Pregnancy on Long-Term Function of Renal Allografts Stephen N. Sturgiss, MB, ChB, MRCOG, and John M. Davison, MD, FRCOG • Pregnancy in renal allograft recipients is associated with hyperfiltration with the potential for glomerular damage and adverse effects on long-term graft prognosis. We have undertaken a case-controlled study of posttransplant follow-up for a mean of 12 years (range, 4 to 23) in 36 female renal allograft recipients, 18 who became pregnant and 18 controls (matched to underlying disease and renal function) who did not. Assessments included plasma creatinine (PCr), glomerular filtration rate (GFR) by infusion clearance of inulin (Cin), mean arterial pressure (MAP), and documentation of antihypertensive therapy. By the end of follow-up, PCr in the pregnancy group (112 ± 73 /Lmol/L [1.26 ± 0.83 mg/dL)) and controls (127 ± 52 /Lmol/L [1.44 ± 0.59 mg/dL)) had increased by 19% and 8%, respectively, and GFR in the pregnancy group (58 ± 29 mL/min) and controls (56 ± 32 mL/min) had decreased by 18% and 7%, respectively. Graft loss or chronic rejection occurred in two patients in each group and there was a death in the pregnancy group 9 years after the second of two successful pregnancies. MAP in the pregnancy group (96 ± 12 mm Hg) had decreased by 1%, and in the controls (101 ± 9 mm Hg) had increased by 5%. Two patients in the index group and three in the control group commenced antihypertensive therapy during follow-up. There was, therefore, no evidence of an adverse effect of pregnancy in renal allograft recipients on long-term renal function or development of hypertension. © 1992 by the National Kidney Foundation, Inc. INDEX WORDS: Pregnancy; renal allograft; hyperfiltration; long-term prognosis.
A
FTER KIDNEY transplantation, renal, endocrine, and sexual function rapidly return and about one in 50 women of childbearing age with a functioning graft becomes pregnant. l Increased glomerular filtration rate (GFR), characteristic of normal pregnancy, is also seen in transplant patients, as is a decrease toward nonpregnant values in the third trimester. 2 This latter change in the majority of cases is not accompanied by other evidence of altered kidney function, nor is it followed by permanent renal impairment. Nevertheless, the impact of pregnancy is important, because after transplantation hyperfiltration in the donated kidney (if associated with increased glomerular or capillary pressure) may cause progressive loss of renal function due to glomerular sclerosis and, theoretically, the additional hyperfiltration of pregnancy could further jeopardize the graft. Due to a paucity of data, statements regarding the effect of pregnancy on long-term graft function are vague. From a review of the literature, it has been estimated that 11 % of women with allografts will have new long-term medical problems after pregnancy,3 but it is difficult to know whether such problems are merely time-related or precipitated by pregnancy. One small study examined follow-up for an average of 7 years in recipients of cadaveric grafts, only five of whom became pregnant, while 10 did not; although matching was imperfect, it was nevertheless concluded that pregnancy had no effect on graft survival or function. 4 Similar conclusions were reached by a European Dialysis and Transplant
Association postal survey that retrospectively analyzed graft function in 53 women who had a successful pregnancy. However, the data from this study are oflimited value, because the longest follow-up was 35 months (with only 15 women available at 15 months beyond delivery), there were no data available for unsuccessful pregnancies, and some of the controls were male transplant recipients. 5 To properly assess the long-term effects of pregnancy on renal allograft function, we have examined remote outcome over an average period of 12 years in 36 female recipients, 18 who became pregnant and 18 who did not, matched at outset for all factors affecting long-term prognosis. MATERIALS AND METHODS A total of 153 women of childbearing age (15 to 35 years) received a renal allograft within the Newcastle Renal Transplant Programme between 1967 and 1987. Table I shows in 3 yearly intervals the number of recipients, graft survival beyond 2 years, and number of women in each cohort subsequently becoming pregnant (mean transplant to pregnancy interval, 6 years; range, I to II). From the Department ojObstetrics and Gynecology, Princess Mary Maternity Hospital, and the Departments oj Medicine and Surgery, Royal Victoria Infirmary and Freeman Road Hospital, University oJNewcastle-upon-Tyne, UK. Supported by the Medical Research Council and Newcastle Health Authority. Address reprint requests to Stephen N. Sturgiss, MB, ChB, MRCOG, Princess Mary Maternity Hospital, Great North Rd, Newcastle-upon-Tyne, NE23BD. UK. © 1992 by the National Kidney Foundation. Inc. 0272-6386/92/1902-0010$3.00/0
American Journal of Kidney Diseases, Vol XIX, No 2 (February), 1992: pp 167-172
167
STURGISS AND DAVISON
168 Table 1. Newcastle Renal Transplant Programme (1967 to 1987): Number of Women of Reproductive Age Who Received Renal Allografts, Had Graft Survival Beyond 2 Years, and Who Subsequently Became Pregnant
Transplanted Women Aged 15 to 35 Years Graft Survival Year
Total No. of Recipients
1967-1969 1970-1972 1973-1975 1976-1978 1979-1981 1982-1984 1985-1987
10 20 20 23 23 25 32
Total
153
Beyond 2
Years (No.)
Women Becoming Pregnant (No.)'
4 11 7 13 18 18 25
1 1 1 4 8 2
96 (63%)
18 (12%)
* Pregnancies occurred between 1 and 11 years posttransplantation
All potential mothers were screened and counseled according to established prepregnancy assessment criteria. 6 Prepregnancy mean arterial pressure (MAP) in all patients was 107 mm Hg (equivalent to 140/90 mm Hg) or less, although four women (eight pregnancies) were receiving antihypertensive medication. All had satisfactory renal function as evidenced by 24-hour creatinine clearance (CCr) greater than 0.67 mL/ s (40 mL/min), plasma creatinine (PCr) less than 150 /lmol/ L (1.70 mgjdL), and 24-hour protein excretion less than 300 mg. Using the transplant occasion as the reference point, acontrol subject was identified for each index subject, matched for date of transplantation (within 6 months of the corresponding index subject), age, and other factors influencing long-term renal prognosis. Some of the controls were in fact women who had been screened as potential mothers but decided against pregnancy. Table 2 itemizes relative frequencies oflong-term prognostic factors for each group. Original renal disease categories in index subjects were chronic pyelonephritis (eight), glomerulonephritis (three), diabetes (two), congenital (two), hypertensive nephropathy (one), urolithiasis (one), and hemolytic-uremic syndrome (one). In controls, the disease categories were chronic pyelonephritis (nine), glomerulonephritis (two), diabetes (one), hypertensive nephropathy (one), hemolytic-uremic syndrome (one), IgA nephropathy (one), systemic lupus erythematosis (one), and congenital (one). In one control patient, an etiology was not established despite a renal biopsy. All patients were taking immunosuppressive drugs: prednisolone 10 mgjd or less, and azathioprine 2 mgjkg bodyweight or less. Posttransplant assessments, undertaken periodically over a mean period of 12 years (range, 4 to 23; 1967 to April 1991) have included for all cases PCr (and reciprocal PCr, [PCr-I]) and measurement of blood pressure. For the index group, mean follow-up after the first pregnancy was 5 years (range,
1 to 14). For 12 cases in each group, GFR (by infusion clearance of inulin [Cin]) was determined at intervals during followup. Ccr follow-up data were not available, as the renal group at Newcastle believe the results could be potentially misleading. Details of infusion protocols and all laboratory methodology have been described elsewhere. 2•7•8 Urinary protein determination was by Bio-Rad Protein Assay (Bio-Rad Laboratories, Richmond, CAl. Prepregnancy values (or time-matched equivalents in controls) were compared with values in 1991 . In cases of graft loss (two in each group) or patient death (one index case), values were taken immediately before that event. The results are presented as means ± SO. The significance of differences between pairs of observations has been assessed by Student's t test and P < 0.05 is reported as statistically significant.
RESULTS
Obstetric Outcome
In the index group, there were 34 pregnancies among 18 women. Ten pregnancies did not go beyond the first trimester: four spontaneous and five therapeutic abortions and one ectopic pregnancy. Of the 24 pregnancies in 17 women that went beyond early pregnancy, 11 had an adverse perinatal outcome: five growth-retarded babies (< 10th percentile), five stillbirths, and one neonatal death. Mean gestation at delivery was 35.2 weeks (range, 27 to 39), with 11 (46%) infants Table 2. Subject Information and Prognostic Factors for Allograft Receipients
Allograft Recipients
Prognostic Factor Age at transplant (yr)* Time since transplant (yr)* Donor status Cadaver (no.) Uving (no.) Early rejection None (no.) 1 episode (no.) 2+ episodes (no.) Primary renal function Immediate (no.) Delayed (no.) HLA match 1 (no.) 2 (no.) 3 (no.) 4+ (no.) * Mean ± SD.
Becoming Pregnant (Index Subjects)
Not Becoming Pregnant (Control Subjects)
22 ± 5 11 ± 5
26 ± 8 11 ± 4
11 7
13 5
6 5 7
4 7 7
15 3
13 5
4 6 4 4
2 6 5 5
169
RENAL ALLOGRAFTS AND PREGNANCY
born before term and five (21 %) at or before 32 weeks' gestation. Mean birthweight (2,404 g; range, 470 to 3,895) was appropriate for gestational age. The indications for the five antepartum cesarean sections were severe hypertension (one), deteriorating renal function (two), cardiotocographic fetal distress (one), and a previous cesarean section (one). The indications for the three intrapartum cesarean sections were cardiotocographic fetal distress (one), failed induction oflabor (one), and cephalopelvic disproportion (one).
Renal Function At the outset, PCr in the index group (94 ± 26 tLmol/L [1.06 ± 0.29 mg/dLD was less than in controls (118 ± 31 tLmol/L [1.33 ± 0.35 mg/dLD (P < 0.05), but by the end offollow-up index PCr (112 ± 73 tLmol/L [1.26 ± 0.83 mg/dLD and control PCr (127 ± 52 tLmol/L [1.44 ± 0.59 mg/ dLD, having increased by 19% and 8%, respectively, were not significantly different (Fig 1). [PCr-l] decreased by 9% in the index group (0.0114 ± 0.003 to 0.0104 ± 0.003) and by 2% in controls (0.009 ± 0.002 to 0.0088 ± 0.003) (Fig 2), neither change being significant over time. 100 80
C
]
!'" CI:
'C"'
60 40
20 0
:3'
e
3
---'0 0.11 E ---= e
...
0.10
....
~
U
0.09
~-------------------------~
~
o
2
4
6
8
10
Follow-up (years)
12
Fig 2. Change in [PCr-'] (1/~mol/L) over 12 years follow-up in subjects (0 0) and controls (e----e). Values are means.
By the end of follow-up, the decrement in GFR (Cin) was 19% in index subjects (71 ± 26 mL/ min to 58 ± 29 mL/min) and 8% in controls (60 ± 21 mL/min to 56 ± 32 mL/min) (Fig 1), the percentage changes being remarkably similar to the alterations in PCr. Chronic allograft rejection was diagnosed in one index patient (but no controls) who had a cadaver donor transplantation in 1978 and a successful pregnancy in 1980. Renal function was satisfactory throughout pregnancy, but PCr has increased from 260 to 378 tLmol/L (2.94 to 4.28 mg/dL) between 1988 and 1990 without a specific identifiable cause. Three control patients currently have PCr values greater than 150 tLmol/L (1.70 mg/dL), but the level is stable in two of these and deterioration of function in the other is thought to be due to extensive nephrocalcinosis.
Graft Loss and Mortality
200
::::. Q
0.12
100
..
2-
u
CI.
0
INDEX SUBJECTS
CONTROL SUBJECTS
Fig 1. (A) GFR (mL/min) by Cin and (8) PCr (~mol/L) at outset (light shade) and at end of follow-up (dark shade) for subjects and controls. Values are means ±SD.
There was one death in the index group (but none in the controls) due to a cerebrovascular accident 15 years after living donor transplantation (and 9 years after the second of two successful pregnancies), at a time when allograft function was satisfactory. One graft loss occurred in the index group and two in the controls. In the index case, conception occurred 8 years after living donor transplantation, and renal function, moderately impaired prepregnancy (PCr, 142 tLmol/L [1.61 mg/dL]; CCr, 0.67 mL/s [40 mL/ min D, was satisfactorily maintained until 31 weeks, when PCr increased to 230 tLmol/L (2.60 mgjdL) and CCrdecreased to 0.25 mL/s [15 mL/
170
STURGISS AND DAVISON
min]. Allograft function only partially recovered postpartum and transplant nephrectomy was performed 11 months later because of advanced chronic rejection. In the control cases, chronic rejection led to return to dialysis 3 and 6 years after cadaver transplantation. Excluding the one death and three allograft losses, antirejection therapy was reduced or maintained at the same level in all other index cases and controls. Hypertension
At the outset, MAP was similar in index cases (97 ± 7 mm Hg) and controls (96 ± 14 mm Hg) with four index patients and seven controls taking antihypertensive therapy. At end of follow-up, MAP in the index cases (96 ± 12 mm Hg) had decreased 1%, and in controls had increased 5% (101 ± 9 mm Hg), but there was no significant difference between these final follow-up values. During follow-up, two additional index cases and three controls started antihypertensive therapy. During the 24 pregnancies that went beyond the first trimester, significant hypertension (MAP > 107 mm Hg) occurred in 17 (71 %) subjects, appearing before 28 weeks in eight. Hypertension was rarely associated with renal dysfunction, but eight of the 16 developed proteinuria (>300 mg/ 24 h). The transplant to pregnancy interval, donor source, number of pregnancies beyond 28 weeks, and the presence or absence of hypertension before the 28th gestational week had no significant effect on long-term alterations of PCr or MAP. DISCUSSION
The ultimate measure of transplant success is long-term patient and graft survival. As it is but 30 years since renal transplantation became widely used in the management of end-stage renal failure, there are few long-term data from sufficiently large series from which to draw conclusions. Furthermore, it must be emphasized that long-term results relate to a period when several aspects of management would be unacceptable by present day standards. Survival figures oflarge numbers of patients worldwide indicate that 70% to 80% of recipients with kidneys from living related donors are alive 5 years after transplanta-
tion, and with cadaver kidneys the figure is 40% to 50%.9.12 Five-year survival is increased if renal function is normal 2 years posttransplant, and patients are advised that this is probably the optimal time to contemplate pregnancy. Most patients will heed this advice and a large proportion of the women who have received renal allografts in Newcastle recently have delayed becoming pregnant (Table 1). Furthermore, all were receiving azathioprine/steroid immunosuppression, and obviously further evaluation will be necessary with the advent of cyclosporine A. Perinatal outcome was disappointing by our earlier standards2,3.6 and that of other series. 13-15 However, the women here comprise a high-risk obstetric population by any standards, as well as by being renal transplant recipients. Eleven of the 24 pregnancies were to women who were either diabetic, receiving antihypertensive medication before pregnancy, and/or were on the borderline between mild and moderate renal insufficiency (PCr, 100 to 160 JLmol/L). Three of the five stillbirths were in the two diabetic women (counseled against becoming pregnant) who were also receiving prepregnancy antihypertensive therapy. Not surprisingly, there was a high incidence of hypertension in our patients, which when present before pregnancy or before 28 weeks, was particularly associated with adverse perinatal outcome. 16 Two further women (not in this study, as long-term postpregnancy follow-up was not available) have just had uncomplicated pregnancies and delivered infants of 2,370 g at 37 weeks and 1,955 g and 1,962 g (twins) at 36 weeks (all three babies being <10th percentile, but otherwise healthy). It has been shown that pregnancy causes an increment in GFR in a renal allograft, which is superimposed on the hyperfiltration already occurring in a single kidney.2 Hyperfiltration in the single-kidney rat model can lead to glomerular sclerosis and declining GFR, which can be partially prevented by alleviating the hyperfiltration with a low-protein diet. 17 Consequently, there is concern that long-term allograft function could be impaired by pregnancy, even though pregnancy hyperfiltration is flow-related without a concomitant increase in intraglomerular pressure in the two-kidney rat model. 18 Indeed, pregnancy is not detrimental in healthy animals as it is glo-
RENAL ALLOGRAFTS AND PREGNANCY
merular hypertension, rather than increased flow, that causes glomerular sclerosis.1 9 The limited data from healthy pregnant women would endorse this view,2 but for the special situation of a single kidney, especially a transplanted one, no reliable information has been available. Exact matching of all long-term prognostic factors is difficult, because renal transplant recipients becoming pregnant will tend to be those with good general health and adequate graft function. However, our groups were adequately matched for prognostic factors including age, early rejection episodes, primary renal function, time since transplant, and the number of 3 or 4+ HLA matches. As well as using clinically applicable measures for renal function-PCr and [PCr-I]-we undertook periodic assessments using infusion Cin in 12 of 18 women in both groups. Index and control PCr increased by 19% and 8%, respectively, a!ld [PCr-l] decreased by 9% and 2%, respectively. These changes over time were nonsignificant. A substantial literature attests to the fact that PCr and/or [PCr-l] can be misleading as indices of renal function in certain circumstances, but our inulin results endorse the creatinine data given here, with 19% and 8% decrements in GFR for index and control cases, respectively. Consequently, this investigation excludes a major deleterious effect of pregnancy on long-term graft function. This lack of an adverse effect is especially encouraging in view of the relatively high risk of the pregnancy group. However, only a small proportion of our population of transplanted women of childbearing age became pregnant, and this, at least in part, reflects our counseling of women against pregnancy where there are poor renal parameters and/or more severe hypertension. Nevertheless, future prepregnancy counseling should continue to be based on established criteria6 if the incidence of adverse long-term outcome is to be kept to acceptable levels. Other stimuli can induce hyperfiltration, such as amino acid infusion or protein meal ingestion (albeit on a smaller scale than pregnancy and almost certainly via a different mechanism), allowing assessment of renal hemodynamic reserve (RHR), another means of quantifying renal function. Using this approach, it has been dem-
171
onstrated l8 that hyperfiltrating midterm pregnant rats still have substantial RHR even at a time of peak gestational increments of GFR and renal plasma flow (RPF). Our unpublished serial inulin and p-aminohippurate clearance data (using an amino acid infusion stimulus) for human pregnancy indicates that, at least in early gestation, RHR is equivalent to that in nonpregnant subjects. We are currently studying renal transplant recipients with each patient acting as her own nonpregnant control to assess whether in this situation RHR has been completely attenuated by pregnancy. There was no significant change in MAP over the study period and comparable numbers of women in the index (two) and the control (three) groups needed to commence antihypertensive medication during follow-up. The cause of hypertension following renal transplantation is often unclear, but it can occur as a consequence of and/or in conjunction with renal dysfunction. Therefore, if graft function was impaired by pregnancy, it might be expected to lead to an increased incidence of hypertension. Our study showed that pregnancy did not affect long-term predisposition to hypertension, further suggesting that pregnancy does not harm graft function. In summary, pregnancy does not seem to have a major effect on long-term allograft survival or function. The decrement in renal function in the index group was greater than in controls and may indicate a minor effect, but this would have to be confirmed with either a larger multicenter study and/or more sensitive indicators of renal function. Pregnancy does not appear to predispose to long-term hypertension, or permanently aggravate hypertension in women already receiving antihypertensive therapy. ACKNOWLEDGMENT The work would not have been possible without the help and support of the Renal Transplant Team at the Royal Victoria Infirmary and Freeman Hospital, Newcastle-upon-Tyne. We are indebted to Professor R. Wilkinson and Drs R. W. Elliott, T. H. J. Goodship, T. W. J. Lennard, G. Proud, J. S. Tapson, R. M. R. Taylor, M. K. Ward, and the late Mr P. J. Dewar.
REFERENCES I. Lau J, Scott JR: Pregnancy following renal transplan-
tation. Clin Obstet Gynecol 28:339-350, 1983
172
2. Davison JM: The effect of pregnancy on kidney function in renal allograft recipients. Kidney Int 27:74-79, 1985 3. Davison JM: Pregnancy in renal allograft recipients: Prognosis and management. Baillieres Clin Obstet Gynaecol 1:1027-1045, 1987 4. Whetham JCG, Cardella C, Harding M: Effect ofpregnancy on graft function and graft survival in renal cadaver transplant patients. Am J Obstet Gynecol 145: 193-197, 1983 5. EDTA: Combined Report on Regular Dialysis and Transplantation in Europe, XIX, 1988. Nephrol Dial Transplant 4:31-40, 1989 . 6. DaviSon JM, Lind T, Uidall PR: Planned pregnancy in a renal transplant recipient. Br J Obstet Gynaecol 83:517527, 1976 7. Goldring W, Chasis W: in Hypertension and Hypertensive Disease. New York, NY, Commonwealth Fund, 1944, p 196 8. Davison JM, Hytten FE: Glomerular filtration rate during and after pregnancy. Br J Obstet Gynaecol 81:558-595, 1974 9. Advisory Committee to the Transplant Registry: The 13th Report of the Human Renal Transplant Registry. Transplant Proc 9:9-26, 1977 10. United Kingdom Transplant Service: Annual Report. Bristol, England, U~ Transplant Service, 1988
STURGISS AND DAVISON
II. Saudie E, Mahoney JF, Catterson RJ: Long-term survival after cadaveric renal transplantation. Br Med J 2: 11601163, 1983 12. Vollmer WM, Wahl PW, Blagg CR: Survival with dialysis and transplantation in patients with end-stage renal disease. N Engl J Med 308:1553-1558, 1983 13. Rifle G, Traeger J: Pregnancy after renal transplantation: An international survey. Transplant Proc 7:723-728, 1975 14. Waltzer WC, Coulam CB, Zincke H, et al: Pregnancy in renal transplantation. Transplant Proc 12:221-226, 1980 15. Penn I, Makowski EL, Harris P: Parenthood following renal transplantation. Kidney Int 18:221-233, 1980 16. Sturgiss SN, Davison JM : Perinatal outcome in renal allograft recipients: Prognostic significance of hypertension and renal function before and during pregnancy. Obstet Gynecol 78:573-577, 1991 17. Cartwright ME, Jaenke RS: Effects of dietary protein and captopril on glomerular permselectivity in rats with unilateral nephrectomy. Lab Invest 59:492-499, 1988 18. Baylis C: Glomerular filtration and volume regulation in gravid animal models. Baillieres Clin Obstet Gynaecol I: 789-813, 1987 19. Anderson S, Brenner BM: The role of intraglomerular pressure in the initiation and progression of renal disease. J Hypertens 4:S236-S238, 1986 (suppl 5)
A
FTER KIDNEY transplantation, renal, endocrine, and sexual function rapidly return and about one in 50 women of childbearing age with a functioning graft becomes pregnant. l Increased glomerular filtration rate (GFR), characteristic of normal pregnancy, is also seen in transplant patients, as is a decrease toward nonpregnant values in the third trimester. 2 This latter change in the majority of cases is not accompanied by other evidence of altered kidney function, nor is it followed by permanent renal impairment. Nevertheless, the impact of pregnancy is important, because after transplantation hyperfiltration in the donated kidney (if associated with increased glomerular or capillary pressure) may cause progressive loss of renal function due to glomerular sclerosis and, theoretically, the additional hyperfiltration of pregnancy could further jeopardize the graft. Due to a paucity of data, statements regarding the effect of pregnancy on long-term graft function are vague. From a review of the literature, it has been estimated that 11 % of women with allografts will have new long-term medical problems after pregnancy,3 but it is difficult to know whether such problems are merely time-related or precipitated by pregnancy. One small study examined follow-up for an average of 7 years in recipients of cadaveric grafts, only five of whom became pregnant, while 10 did not; although matching was imperfect, it was nevertheless concluded that pregnancy had no effect on graft survival or function. 4 Similar conclusions were reached by a European Dialysis and Transplant
Association postal survey that retrospectively analyzed graft function in 53 women who had a successful pregnancy. However, the data from this study are oflimited value, because the longest follow-up was 35 months (with only 15 women available at 15 months beyond delivery), there were no data available for unsuccessful pregnancies, and some of the controls were male transplant recipients. 5 To properly assess the long-term effects of pregnancy on renal allograft function, we have examined remote outcome over an average period of 12 years in 36 female recipients, 18 who became pregnant and 18 who did not, matched at outset for all factors affecting long-term prognosis. MATERIALS AND METHODS A total of 153 women of childbearing age (15 to 35 years) received a renal allograft within the Newcastle Renal Transplant Programme between 1967 and 1987. Table I shows in 3 yearly intervals the number of recipients, graft survival beyond 2 years, and number of women in each cohort subsequently becoming pregnant (mean transplant to pregnancy interval, 6 years; range, I to II). From the Department ojObstetrics and Gynecology, Princess Mary Maternity Hospital, and the Departments oj Medicine and Surgery, Royal Victoria Infirmary and Freeman Road Hospital, University oJNewcastle-upon-Tyne, UK. Supported by the Medical Research Council and Newcastle Health Authority. Address reprint requests to Stephen N. Sturgiss, MB, ChB, MRCOG, Princess Mary Maternity Hospital, Great North Rd, Newcastle-upon-Tyne, NE23BD. UK. © 1992 by the National Kidney Foundation. Inc. 0272-6386/92/1902-0010$3.00/0
American Journal of Kidney Diseases, Vol XIX, No 2 (February), 1992: pp 167-172
167
STURGISS AND DAVISON
168 Table 1. Newcastle Renal Transplant Programme (1967 to 1987): Number of Women of Reproductive Age Who Received Renal Allografts, Had Graft Survival Beyond 2 Years, and Who Subsequently Became Pregnant
Transplanted Women Aged 15 to 35 Years Graft Survival Year
Total No. of Recipients
1967-1969 1970-1972 1973-1975 1976-1978 1979-1981 1982-1984 1985-1987
10 20 20 23 23 25 32
Total
153
Beyond 2
Years (No.)
Women Becoming Pregnant (No.)'
4 11 7 13 18 18 25
1 1 1 4 8 2
96 (63%)
18 (12%)
* Pregnancies occurred between 1 and 11 years posttransplantation
All potential mothers were screened and counseled according to established prepregnancy assessment criteria. 6 Prepregnancy mean arterial pressure (MAP) in all patients was 107 mm Hg (equivalent to 140/90 mm Hg) or less, although four women (eight pregnancies) were receiving antihypertensive medication. All had satisfactory renal function as evidenced by 24-hour creatinine clearance (CCr) greater than 0.67 mL/ s (40 mL/min), plasma creatinine (PCr) less than 150 /lmol/ L (1.70 mgjdL), and 24-hour protein excretion less than 300 mg. Using the transplant occasion as the reference point, acontrol subject was identified for each index subject, matched for date of transplantation (within 6 months of the corresponding index subject), age, and other factors influencing long-term renal prognosis. Some of the controls were in fact women who had been screened as potential mothers but decided against pregnancy. Table 2 itemizes relative frequencies oflong-term prognostic factors for each group. Original renal disease categories in index subjects were chronic pyelonephritis (eight), glomerulonephritis (three), diabetes (two), congenital (two), hypertensive nephropathy (one), urolithiasis (one), and hemolytic-uremic syndrome (one). In controls, the disease categories were chronic pyelonephritis (nine), glomerulonephritis (two), diabetes (one), hypertensive nephropathy (one), hemolytic-uremic syndrome (one), IgA nephropathy (one), systemic lupus erythematosis (one), and congenital (one). In one control patient, an etiology was not established despite a renal biopsy. All patients were taking immunosuppressive drugs: prednisolone 10 mgjd or less, and azathioprine 2 mgjkg bodyweight or less. Posttransplant assessments, undertaken periodically over a mean period of 12 years (range, 4 to 23; 1967 to April 1991) have included for all cases PCr (and reciprocal PCr, [PCr-I]) and measurement of blood pressure. For the index group, mean follow-up after the first pregnancy was 5 years (range,
1 to 14). For 12 cases in each group, GFR (by infusion clearance of inulin [Cin]) was determined at intervals during followup. Ccr follow-up data were not available, as the renal group at Newcastle believe the results could be potentially misleading. Details of infusion protocols and all laboratory methodology have been described elsewhere. 2•7•8 Urinary protein determination was by Bio-Rad Protein Assay (Bio-Rad Laboratories, Richmond, CAl. Prepregnancy values (or time-matched equivalents in controls) were compared with values in 1991 . In cases of graft loss (two in each group) or patient death (one index case), values were taken immediately before that event. The results are presented as means ± SO. The significance of differences between pairs of observations has been assessed by Student's t test and P < 0.05 is reported as statistically significant.
RESULTS
Obstetric Outcome
In the index group, there were 34 pregnancies among 18 women. Ten pregnancies did not go beyond the first trimester: four spontaneous and five therapeutic abortions and one ectopic pregnancy. Of the 24 pregnancies in 17 women that went beyond early pregnancy, 11 had an adverse perinatal outcome: five growth-retarded babies (< 10th percentile), five stillbirths, and one neonatal death. Mean gestation at delivery was 35.2 weeks (range, 27 to 39), with 11 (46%) infants Table 2. Subject Information and Prognostic Factors for Allograft Receipients
Allograft Recipients
Prognostic Factor Age at transplant (yr)* Time since transplant (yr)* Donor status Cadaver (no.) Uving (no.) Early rejection None (no.) 1 episode (no.) 2+ episodes (no.) Primary renal function Immediate (no.) Delayed (no.) HLA match 1 (no.) 2 (no.) 3 (no.) 4+ (no.) * Mean ± SD.
Becoming Pregnant (Index Subjects)
Not Becoming Pregnant (Control Subjects)
22 ± 5 11 ± 5
26 ± 8 11 ± 4
11 7
13 5
6 5 7
4 7 7
15 3
13 5
4 6 4 4
2 6 5 5
169
RENAL ALLOGRAFTS AND PREGNANCY
born before term and five (21 %) at or before 32 weeks' gestation. Mean birthweight (2,404 g; range, 470 to 3,895) was appropriate for gestational age. The indications for the five antepartum cesarean sections were severe hypertension (one), deteriorating renal function (two), cardiotocographic fetal distress (one), and a previous cesarean section (one). The indications for the three intrapartum cesarean sections were cardiotocographic fetal distress (one), failed induction oflabor (one), and cephalopelvic disproportion (one).
Renal Function At the outset, PCr in the index group (94 ± 26 tLmol/L [1.06 ± 0.29 mg/dLD was less than in controls (118 ± 31 tLmol/L [1.33 ± 0.35 mg/dLD (P < 0.05), but by the end offollow-up index PCr (112 ± 73 tLmol/L [1.26 ± 0.83 mg/dLD and control PCr (127 ± 52 tLmol/L [1.44 ± 0.59 mg/ dLD, having increased by 19% and 8%, respectively, were not significantly different (Fig 1). [PCr-l] decreased by 9% in the index group (0.0114 ± 0.003 to 0.0104 ± 0.003) and by 2% in controls (0.009 ± 0.002 to 0.0088 ± 0.003) (Fig 2), neither change being significant over time. 100 80
C
]
!'" CI:
'C"'
60 40
20 0
:3'
e
3
---'0 0.11 E ---= e
...
0.10
....
~
U
0.09
~-------------------------~
~
o
2
4
6
8
10
Follow-up (years)
12
Fig 2. Change in [PCr-'] (1/~mol/L) over 12 years follow-up in subjects (0 0) and controls (e----e). Values are means.
By the end of follow-up, the decrement in GFR (Cin) was 19% in index subjects (71 ± 26 mL/ min to 58 ± 29 mL/min) and 8% in controls (60 ± 21 mL/min to 56 ± 32 mL/min) (Fig 1), the percentage changes being remarkably similar to the alterations in PCr. Chronic allograft rejection was diagnosed in one index patient (but no controls) who had a cadaver donor transplantation in 1978 and a successful pregnancy in 1980. Renal function was satisfactory throughout pregnancy, but PCr has increased from 260 to 378 tLmol/L (2.94 to 4.28 mg/dL) between 1988 and 1990 without a specific identifiable cause. Three control patients currently have PCr values greater than 150 tLmol/L (1.70 mg/dL), but the level is stable in two of these and deterioration of function in the other is thought to be due to extensive nephrocalcinosis.
Graft Loss and Mortality
200
::::. Q
0.12
100
..
2-
u
CI.
0
INDEX SUBJECTS
CONTROL SUBJECTS
Fig 1. (A) GFR (mL/min) by Cin and (8) PCr (~mol/L) at outset (light shade) and at end of follow-up (dark shade) for subjects and controls. Values are means ±SD.
There was one death in the index group (but none in the controls) due to a cerebrovascular accident 15 years after living donor transplantation (and 9 years after the second of two successful pregnancies), at a time when allograft function was satisfactory. One graft loss occurred in the index group and two in the controls. In the index case, conception occurred 8 years after living donor transplantation, and renal function, moderately impaired prepregnancy (PCr, 142 tLmol/L [1.61 mg/dL]; CCr, 0.67 mL/s [40 mL/ min D, was satisfactorily maintained until 31 weeks, when PCr increased to 230 tLmol/L (2.60 mgjdL) and CCrdecreased to 0.25 mL/s [15 mL/
170
STURGISS AND DAVISON
min]. Allograft function only partially recovered postpartum and transplant nephrectomy was performed 11 months later because of advanced chronic rejection. In the control cases, chronic rejection led to return to dialysis 3 and 6 years after cadaver transplantation. Excluding the one death and three allograft losses, antirejection therapy was reduced or maintained at the same level in all other index cases and controls. Hypertension
At the outset, MAP was similar in index cases (97 ± 7 mm Hg) and controls (96 ± 14 mm Hg) with four index patients and seven controls taking antihypertensive therapy. At end of follow-up, MAP in the index cases (96 ± 12 mm Hg) had decreased 1%, and in controls had increased 5% (101 ± 9 mm Hg), but there was no significant difference between these final follow-up values. During follow-up, two additional index cases and three controls started antihypertensive therapy. During the 24 pregnancies that went beyond the first trimester, significant hypertension (MAP > 107 mm Hg) occurred in 17 (71 %) subjects, appearing before 28 weeks in eight. Hypertension was rarely associated with renal dysfunction, but eight of the 16 developed proteinuria (>300 mg/ 24 h). The transplant to pregnancy interval, donor source, number of pregnancies beyond 28 weeks, and the presence or absence of hypertension before the 28th gestational week had no significant effect on long-term alterations of PCr or MAP. DISCUSSION
The ultimate measure of transplant success is long-term patient and graft survival. As it is but 30 years since renal transplantation became widely used in the management of end-stage renal failure, there are few long-term data from sufficiently large series from which to draw conclusions. Furthermore, it must be emphasized that long-term results relate to a period when several aspects of management would be unacceptable by present day standards. Survival figures oflarge numbers of patients worldwide indicate that 70% to 80% of recipients with kidneys from living related donors are alive 5 years after transplanta-
tion, and with cadaver kidneys the figure is 40% to 50%.9.12 Five-year survival is increased if renal function is normal 2 years posttransplant, and patients are advised that this is probably the optimal time to contemplate pregnancy. Most patients will heed this advice and a large proportion of the women who have received renal allografts in Newcastle recently have delayed becoming pregnant (Table 1). Furthermore, all were receiving azathioprine/steroid immunosuppression, and obviously further evaluation will be necessary with the advent of cyclosporine A. Perinatal outcome was disappointing by our earlier standards2,3.6 and that of other series. 13-15 However, the women here comprise a high-risk obstetric population by any standards, as well as by being renal transplant recipients. Eleven of the 24 pregnancies were to women who were either diabetic, receiving antihypertensive medication before pregnancy, and/or were on the borderline between mild and moderate renal insufficiency (PCr, 100 to 160 JLmol/L). Three of the five stillbirths were in the two diabetic women (counseled against becoming pregnant) who were also receiving prepregnancy antihypertensive therapy. Not surprisingly, there was a high incidence of hypertension in our patients, which when present before pregnancy or before 28 weeks, was particularly associated with adverse perinatal outcome. 16 Two further women (not in this study, as long-term postpregnancy follow-up was not available) have just had uncomplicated pregnancies and delivered infants of 2,370 g at 37 weeks and 1,955 g and 1,962 g (twins) at 36 weeks (all three babies being <10th percentile, but otherwise healthy). It has been shown that pregnancy causes an increment in GFR in a renal allograft, which is superimposed on the hyperfiltration already occurring in a single kidney.2 Hyperfiltration in the single-kidney rat model can lead to glomerular sclerosis and declining GFR, which can be partially prevented by alleviating the hyperfiltration with a low-protein diet. 17 Consequently, there is concern that long-term allograft function could be impaired by pregnancy, even though pregnancy hyperfiltration is flow-related without a concomitant increase in intraglomerular pressure in the two-kidney rat model. 18 Indeed, pregnancy is not detrimental in healthy animals as it is glo-
RENAL ALLOGRAFTS AND PREGNANCY
merular hypertension, rather than increased flow, that causes glomerular sclerosis.1 9 The limited data from healthy pregnant women would endorse this view,2 but for the special situation of a single kidney, especially a transplanted one, no reliable information has been available. Exact matching of all long-term prognostic factors is difficult, because renal transplant recipients becoming pregnant will tend to be those with good general health and adequate graft function. However, our groups were adequately matched for prognostic factors including age, early rejection episodes, primary renal function, time since transplant, and the number of 3 or 4+ HLA matches. As well as using clinically applicable measures for renal function-PCr and [PCr-I]-we undertook periodic assessments using infusion Cin in 12 of 18 women in both groups. Index and control PCr increased by 19% and 8%, respectively, a!ld [PCr-l] decreased by 9% and 2%, respectively. These changes over time were nonsignificant. A substantial literature attests to the fact that PCr and/or [PCr-l] can be misleading as indices of renal function in certain circumstances, but our inulin results endorse the creatinine data given here, with 19% and 8% decrements in GFR for index and control cases, respectively. Consequently, this investigation excludes a major deleterious effect of pregnancy on long-term graft function. This lack of an adverse effect is especially encouraging in view of the relatively high risk of the pregnancy group. However, only a small proportion of our population of transplanted women of childbearing age became pregnant, and this, at least in part, reflects our counseling of women against pregnancy where there are poor renal parameters and/or more severe hypertension. Nevertheless, future prepregnancy counseling should continue to be based on established criteria6 if the incidence of adverse long-term outcome is to be kept to acceptable levels. Other stimuli can induce hyperfiltration, such as amino acid infusion or protein meal ingestion (albeit on a smaller scale than pregnancy and almost certainly via a different mechanism), allowing assessment of renal hemodynamic reserve (RHR), another means of quantifying renal function. Using this approach, it has been dem-
171
onstrated l8 that hyperfiltrating midterm pregnant rats still have substantial RHR even at a time of peak gestational increments of GFR and renal plasma flow (RPF). Our unpublished serial inulin and p-aminohippurate clearance data (using an amino acid infusion stimulus) for human pregnancy indicates that, at least in early gestation, RHR is equivalent to that in nonpregnant subjects. We are currently studying renal transplant recipients with each patient acting as her own nonpregnant control to assess whether in this situation RHR has been completely attenuated by pregnancy. There was no significant change in MAP over the study period and comparable numbers of women in the index (two) and the control (three) groups needed to commence antihypertensive medication during follow-up. The cause of hypertension following renal transplantation is often unclear, but it can occur as a consequence of and/or in conjunction with renal dysfunction. Therefore, if graft function was impaired by pregnancy, it might be expected to lead to an increased incidence of hypertension. Our study showed that pregnancy did not affect long-term predisposition to hypertension, further suggesting that pregnancy does not harm graft function. In summary, pregnancy does not seem to have a major effect on long-term allograft survival or function. The decrement in renal function in the index group was greater than in controls and may indicate a minor effect, but this would have to be confirmed with either a larger multicenter study and/or more sensitive indicators of renal function. Pregnancy does not appear to predispose to long-term hypertension, or permanently aggravate hypertension in women already receiving antihypertensive therapy. ACKNOWLEDGMENT The work would not have been possible without the help and support of the Renal Transplant Team at the Royal Victoria Infirmary and Freeman Hospital, Newcastle-upon-Tyne. We are indebted to Professor R. Wilkinson and Drs R. W. Elliott, T. H. J. Goodship, T. W. J. Lennard, G. Proud, J. S. Tapson, R. M. R. Taylor, M. K. Ward, and the late Mr P. J. Dewar.
REFERENCES I. Lau J, Scott JR: Pregnancy following renal transplan-
tation. Clin Obstet Gynecol 28:339-350, 1983
172
2. Davison JM: The effect of pregnancy on kidney function in renal allograft recipients. Kidney Int 27:74-79, 1985 3. Davison JM: Pregnancy in renal allograft recipients: Prognosis and management. Baillieres Clin Obstet Gynaecol 1:1027-1045, 1987 4. Whetham JCG, Cardella C, Harding M: Effect ofpregnancy on graft function and graft survival in renal cadaver transplant patients. Am J Obstet Gynecol 145: 193-197, 1983 5. EDTA: Combined Report on Regular Dialysis and Transplantation in Europe, XIX, 1988. Nephrol Dial Transplant 4:31-40, 1989 . 6. DaviSon JM, Lind T, Uidall PR: Planned pregnancy in a renal transplant recipient. Br J Obstet Gynaecol 83:517527, 1976 7. Goldring W, Chasis W: in Hypertension and Hypertensive Disease. New York, NY, Commonwealth Fund, 1944, p 196 8. Davison JM, Hytten FE: Glomerular filtration rate during and after pregnancy. Br J Obstet Gynaecol 81:558-595, 1974 9. Advisory Committee to the Transplant Registry: The 13th Report of the Human Renal Transplant Registry. Transplant Proc 9:9-26, 1977 10. United Kingdom Transplant Service: Annual Report. Bristol, England, U~ Transplant Service, 1988
STURGISS AND DAVISON
II. Saudie E, Mahoney JF, Catterson RJ: Long-term survival after cadaveric renal transplantation. Br Med J 2: 11601163, 1983 12. Vollmer WM, Wahl PW, Blagg CR: Survival with dialysis and transplantation in patients with end-stage renal disease. N Engl J Med 308:1553-1558, 1983 13. Rifle G, Traeger J: Pregnancy after renal transplantation: An international survey. Transplant Proc 7:723-728, 1975 14. Waltzer WC, Coulam CB, Zincke H, et al: Pregnancy in renal transplantation. Transplant Proc 12:221-226, 1980 15. Penn I, Makowski EL, Harris P: Parenthood following renal transplantation. Kidney Int 18:221-233, 1980 16. Sturgiss SN, Davison JM : Perinatal outcome in renal allograft recipients: Prognostic significance of hypertension and renal function before and during pregnancy. Obstet Gynecol 78:573-577, 1991 17. Cartwright ME, Jaenke RS: Effects of dietary protein and captopril on glomerular permselectivity in rats with unilateral nephrectomy. Lab Invest 59:492-499, 1988 18. Baylis C: Glomerular filtration and volume regulation in gravid animal models. Baillieres Clin Obstet Gynaecol I: 789-813, 1987 19. Anderson S, Brenner BM: The role of intraglomerular pressure in the initiation and progression of renal disease. J Hypertens 4:S236-S238, 1986 (suppl 5)