Comparison of clinical features of pregnant and non-pregnant females with primary hyperoxaluria

Primary hyperoxaluria (PH) is a rare monogenic disorder of glyoxylate metabolism with three phenotypes [1]. PH1 accounts for 70–80% of all diagnosed cases; PH2 and PH3 each account for approximately 10–15%. Increased oxalate production leads to hyperoxaluria, recurrent urolithiasis, and accumulation of oxalate in the kidneys, which can result in chronic kidney disease and eventually kidney failure [2]. Information on maternal and infant outcomes from pregnancy in women with PH is limited, and data about disease progression in women with PH in relation to pregnancy are lacking.

In this study, we evaluated infant and maternal complications of pregnancy in women with PH and characterized changes in kidney function as a consequence of pregnancy. We also compared the clinical features and risk of kidney failure in PH women with a history of pregnancy to those without pregnancy. We identified women age 20 and older in the Rare Kidney Stone Consortium (RKSC) from January 1, 2003 to November 30, 2020, reviewed available medical records, and sent an email to all women requesting a phone interview. Overall, we identified 46 females who had at least one pregnancy and 38 females who did not have pregnancy (Figure S1).

Phone interviews provided detailed information on 37 pregnancies in 15 unique females to supplement information available in the medical records. There were 29 live births (78.4% compared to general population livebirth rate with clinically recognized pregnancy approximately 75% [3]) with 27 term and 2 preterm deliveries and an average birthweight of 3.4±0.7 kg (Table S1). Approximately 30% of the pregnancies were reported to have no maternal complications. Flank pain and/or hematuria were the most common complications (Table S2) reported during pregnancy by 57% of all PH patients, and 78% of females with PH1. Among the 29 live-born infants, 79% of infants and 54% of children had no medical conditions (Table S2). As reported previously [4], our data also suggest that pregnancy appeared to be well tolerated by women with PH.

By combining our current phone interview results with an earlier phone survey study in a similar RKSC cohort (ending in the year 2000) [4] we had information on timing of pregnancy in relation to overall disease course in 29 women. Figure 1 demonstrates the time course of the diagnosis of PH, pregnancy, and kidney failure in these 29 women. None developed kidney failure during pregnancy. Twelve (41%) women, all with PH1, progressed to kidney failure 14 years (interquartile [IQR] 3.7, 30.8) after their first delivery. In the 7 women with multiple pregnancies, kidney failure occurred 18 years (IQR 12, 26.3) after the last delivery. Eight (28%) of the females were diagnosed 24 years (IQR 14.7, 31.5) prior to the first pregnancy, while 21 (72%) females were diagnosed 9 years (IQR 4.9, 23) after their first delivery. Notably, only two patients, both with PH1, developed kidney failure within 1 year after their first delivery (patients #1 and #20); one of whom underwent combined liver-kidney transplantation (16 months later) and the other kidney transplant alone (8 months later). One of these patients experienced severe hyperemesis gravidarum. Patient #20 went on to have 3 pregnancies after kidney transplantation. Patient #2 was the only woman in the cohort who had all of her pregnancies after transplant. The other 9 females, all with PH1, received either kidney alone or combined liver-kidney transplant a median of 12 years (IQR 7.6, 26.3) after the last delivery.

Figure 1.

Time from the first and last pregnancy to diagnosis of primary hyperoxaluria, kidney failure and transplantation.

^a Referent was the delivery date of the first pregnancy. Patients were ordered by PH type and the time of diagnosis of PH relative to the first pregnancy. The eGFR was calculated by CKD-EPI. Kidney failure was defined as an eGFR < 15 ml/min per 1.73 m², or the need for dialysis or transplant.

We did not have serial measures of kidney function in all women, but did have several decades of follow-up in some cases and did not see any significant change in the median eGFR or 24-hour urine oxalate excretion (U_[ox]) around the time of pregnancy (Figure S2A, B). Paired t test confirmed that eGFR did not change significantly in six women who had a measurement within one year prior to pregnancy and between six months and one year after delivery (Figure S2C). We compared clinical features and the risk of kidney failure in females with PH and a history of pregnancy (n=46) to those without pregnancy (n=38) by combining all available data from medical records and phone surveys. Women who had pregnancies were older at diagnosis and demonstrated a longer interval between symptoms onset and diagnosis; however, there was no significant difference in other clinical characteristics including PH type, eGFR and U_[ox] at diagnosis, and the rate of kidney failure between the two groups (Table S3; Figure S3). Kaplan-Meier analysis demonstrated kidney survival estimates to be similar between patients with pregnancy and those without pregnancy (Figure S4), and Cox model demonstrated that pregnancy did not seem to be associated with kidney failure when adjusted for PH type, age, eGFR and U_[ox] at diagnosis (Table S4). Logistic regression analysis revealed that pregnancy was not a significant kidney failure risk factor (Table S5). Taken as a whole, our results suggest that pregnancy does not impact the risk of kidney failure or decline in eGFR in most women with PH.

Our results are notable for several reasons. Unlike other etiologies of chronic kidney disease, this case series suggests that women with PH do not appear to experience a decline in kidney function as a consequence of pregnancy. This could be due to changes in protein and urea synthesis as an adaptation to pregnancy that may affect oxalate metabolism [5]. However, this could also be because women with PH in our cohort who had pregnancy had less severe disease and relatively preserved GFR at the time of conception (Figure S2C). The women with pregnancies tended to be diagnosed at an older age than women without pregnancy (median (IQR) years: 32 (18, 41) vs. 13 (5, 23), respectively). Furthermore, although pregnancy is a time of increased health-care utilization for women, the diagnosis of PH did not appear to coincide with the time of pregnancy, despite the presentation of women with flank pain, hematuria, and stone passage. Thus, although PH is rare, providers should maintain a high index of suspicion if women develop these symptoms in pregnancy, particularly if they appear to have a high stone burden in the absence of metabolic risk factors. Given the potential for complications related to stones, PH patients require close monitoring during pregnancy to detect and address potential complications seen in this population.

Our study has several limitations. We did not know about pregnancy status in 40 of the women as since they had no medical records to review and did not participate in the telephone survey. Therefore, the coverage of the entire cohort is sub-optimal, with only 68% (84/124) registry patients having a confirmed pregnancy status. The sample size, in particular the women with PH2 or PH3, was small likely resulting in low statistical power. Second, the non-pregnancy group was identified and studied by medical records review and therefore mention of pregnancy may not have been included in the records, or the information may not be fully current. The potential for recall bias may exist in a phone interview. Third, this is a retrospective study of a clinical registry, and thus a voluntary dataset that is inherently incomplete. The limited data of continuous eGFR and U_[ox] make it hard to directly assess changes before and after pregnancy. Fourth, the comparison of renal outcomes could be confounded by several factors, such as the large gap in age, disease severity, whether or not the disease was known before pregnancy, and kidney function.

In conclusion, this is the first large cohort study to compare the influence of pregnancy on maternal renal outcomes in females with PH to those without pregnancy. These data suggest that pregnancy generally was not associated with kidney function decline in females with PH, and there does not appear to be an increased risk of maternal and infant complications. Potential fetal and maternal consequences of highly promising siRNA therapeutic agents recently approved for clinical use in PH1 or in late-stage clinical trials are as yet unknown. Women with PH1 may face decisions to interrupt such treatment during pregnancy or to defer pregnancy until further information becomes available. Overall, our results are reassuring for women with PH who want to pursue pregnancy.

Supplementary Material

Supplementary

Supplemental Tables

Table S1. Characteristics of females with primary hyperoxaluria who had pregnancies

Table S2. Maternal and infant complications in primary hyperoxaluria pregnancies

Table S3. Comparison of clinical characteristics of females with primary hyperoxaluria who had a history of pregnancies or those who had never been pregnant

Table S4. Proportional hazards risk of incident kidney failure in females with a history of pregnancies versus never pregnant.

Table S5. Univariate and multivariate analysis of factors associated with kidney failure in the entire cohort.

Supplemental Figures

Figure S1. Patient selection

We identified all women with PH and with a current age greater than 20 who were enrolled in the Rare Kidney Stone Consortium (RKSC) registry between January 1, 2003 and November 30, 2020. Pregnancies were identified in several ways. Sixteen women with a history of pregnancy from a previously published cohort (Norby, S.M. and D.S. Milliner. Am J Kidney Dis. 2004;43:277–85) were registry participants as of 2003 and their outcomes were included. All eligible registry participants were sent an email invitation to participate in a phone interview. If a woman was a member of both cohorts, the most recent responses were used. ^a Two females were reported by previously published cohort (Norby, S.M. and D.S. Milliner. Am J Kidney Dis. 2004;43:277–85).

Figure S2. Change of eGFR and U_[ox] before or after pregnancy in females with PH who had a history of pregnancies.

A. Change of eGFR over 65 years before and after the first pregnancy. Number of patients with measurements at each time point are shown below the x-axis. B. Change of U_[ox] over 50 years before and after the first pregnancy. Number of patients with measurements at each time point are shown below the x-axis. C. Comparison of the most recent eGFR (≤ 1 year) prior to pregnancy and the first eGFR after delivery (at least > 6 months but < 1 year after delivery). Paired t test. n=6. Values of eGFR and 24-hour urine oxalate excretion (U_[ox], mmol/24 h per 1.73 m²) at the time points after transplantation were excluded in females who received transplant.

Figure S3. Time from the date of diagnosis of primary hyperoxaluria to kidney failure and transplant.

Figure S4. Kaplan-Meier plots of renal survival in patients with PH1.

A. Renal survival estimates were assessed among patients with PH1 who did not have kidney failure at the 1^st pregnancy. B. Among patients with PH1 who did not have kidney failure at diagnosis, renal survival estimates did not seem to differ between patients with a history of pregnancies and those who were never pregnant (P=0.6716). C. Overall, among all patients with PH1, renal survival estimates did not seem to differ between patients with a history of pregnancies and those who were never pregnant (P=0.1573).