Pregnancy after living kidney donation, a systematic review of the available evidence, and a review of the current guidance

Abstract Understanding and communicating the risk of pregnancy complications post‐living kidney donation is imperative as the majority of living kidney donors (LKD) are women of childbearing age. We aimed to identify all original research articles examining complications in post‐donation pregnancies and compared the quality and consistency of related guidelines. We searched Embase, MEDLINE, PubMed, society webpages, and guideline registries for English‐language publications published up until December 18, 2020. Ninety‐three articles were screened from which 16 studies were identified, with a total of 1399 post‐donation pregnancies. The outcome of interest, post‐donation pregnancy complications, was not calculable, and only a narrative synthesis of the evidence was possible. The absolute risk of pre‐eclampsia increased from ~1%–3% pre‐donation (lower than the general population) to ~4%–10% post‐donation (comparable to the general population). The risks of adverse fetal and neonatal outcomes were no different between post‐donation and pre‐donation pregnancies. Guidelines and consensus statements were consistent in stating the need to inform LKDs of their post‐donation pregnancy risk, however, the depth and scope of this guidance were variable. While the absolute risk of pregnancy complications remains low post‐donation, a concerted effort is required to better identify and individualize risk in these women, such that consent to donation is truly informed.

Four case studies/series were identified but not discussed further. 20,23,25,28 Risk-of-bias, varied from low 21,24,26 to moderate 12,22,30 to serious [17][18][19] (Figure 2). Studies with a low risk of bias were broadly consistent in the direction of their findings. The certainty in the evidence was deemed low for the outcomes gestational hypertension and pre-eclampsia (Table 2) and very low for gestation (<37 weeks) and birthweight (<2500 g, Table 2). The percentage of hypertensive disorders in pregnancy and fetal complications in the pre-donation, post-donation, and general population are summarized in Figures 3   and 4, respectively. Although all pregnancy complications were included, the predominant complications identified were hypertensive disorders of pregnancy, proteinuria, and gestational diabetes.

| Maternal outcomes
The initial three studies by Buszta et al., 17 Jones et al., 18 and Wrenshall et al., 19 published over 25 years ago, were all US-based single-center retrospective studies, with a study design that was either a review of the medical notes, a telephone-or postal-survey, respectively. The AJT PIPPIAS et al.
LKD pregnancy outcomes in these studies were compared to general population outcomes and not to well-matched control groups. In 39 post-donation pregnancies in 23 LKD studied by Buszta, all women were normotensive during pregnancy. 17 Similarly, none of the 25 pregnancies in 14 post-donation women in the Jones cohort were complicated by gestational hypertension. 18 Wrenshall reported preeclampsia in 4.4% of post-donation pregnancies as compared to the general population incidence of 6%-8%. 19 Based on the poor quality of these studies (i.e., high risk of recall-, survival-and response-bias, small sample sizes, and incomplete data reporting), they were excluded from Figures 3 and 4.
We identified seven retrospective cohort studies with control groups. 11,21,22,24,26,27,30 Of these, three had a before-afterdesign, 21,22,30 whereby pregnancy outcomes were assessed in women both before and after LKD and three studies compared the pregnancy outcomes of LKD with ("healthy") controls. 21 and proteinuria (4.3% vs. 1.1%, p < .0001). Overall, post-donation outcomes were worse than pre-donation, but similar to general population outcomes. Expanding on the same cohort, the group presented the long-term consequences of post-donation pre-eclampsia in the LKDs (followed up for 21 ± 12 years from the index pregnancy) in a 2017 abstract. 11 They reported an increased likelihood of proteinuria, hypertension, and eGFR decline in post-donation pregnancies complicated by pre-eclampsia, though these risks were comparable to the risks experienced by the general population with pre-eclampsia. 11 In a Korean study of 225 LKD, Yoo  Before-and-afterdesign. The occurrence of pre-eclampsia was more common after donation than before (in adjusted analysis only). sub-group was not stated and the observed differences may be confounded by the absence of co-morbidity-matched controls.

| Fetal and neonatal outcomes
In most studies, there were no differences in adverse fetal and neonatal outcomes between post-donation, pre-donation, and non-donor pregnancies ( Figure 4). Reisaeter and Garg found similar rates of low birthweight (<2500 g) or premature (delivery <37 weeks) neonates. 21,24 In the study by Yoo, no neonates from post-donation pregnancies were of a low birthweight, although it should be mentioned that this was defined as <2800 g, as compared to the more widely adopted threshold of <2500 g. 30 There were no fetal or neonatal deaths in the post-donation cohort studied by Garg    The consensus from the expert opinions was that women should be counseled on the increased likelihood of gestational hypertension or pre-eclampsia (Table 2). 39,[41][42][43][44][45][46] Mandelbrot et al. recommended quantifying this increased risk, advising clinicians state that "overall the rates of gestational hypertension or pre-eclampsia increase from ~5% to 11%." 41 However, they continued by recommending that pregnancy care is tailored according to a woman's individualized risk of complications, taking into consideration the impact of race, BMI, and age on their risk of preeclampsia. 41 Lentine and Segev point out that potential LKD must also be informed that most women had uncomplicated pregnancies after kidney donation. 45

| DISCUSS ION
We performed a systematic review to answer the question "Are  Yes. [Exclude those] "with a history of toxemia in pregnancy and whose family is incomplete".
Exclude from donation: "Premenopausal potential donor with a history of toxemia in pregnancy <10 years ago.
[Those with] a history of toxemia in pregnancy in recurrent pregnancies. Premenopausal potential donor with a history of toxemia in pregnancy and whose family is incomplete".
European renal best practice (ERBP); ERBP guideline on kidney donor and recipient evaluation and perioperative care 2015 33 2015 Europe ERA-EDTA We recommend informing women of childbearing age that as they are selected from a very healthy subpopulation, donation increases their individual risk from below that of the general population, to that of the general population (1B)"

(Continues)
PIPPIAS et al. in mind that comparison with the general population may be prone to confounding because live kidney donors are generally considered to be in better health.

Year of publication
Most importantly the effects of donation on maternal and fetal outcomes should be part of the routine discussion about the risks of donation during the informed consent procedure.
Quality of evidence available based on the GRADE tool was reported as poor. The generalizability of the results of these studies is limited to white women, as up to 98% were of this ethnicity ( Table 1).
Information on post-donation risk of hypertensive complications of pregnancy in ethnic minority groups (EMG) cannot be inferred from the studies to date. Women from EMG are known to have a higher risk of pre-eclampsia. 48 Moreover, a recent review suggests that African American women with pre-eclampsia experience more severe hypertension and increased mortality as compared to women of other ethnicities. 49 As such, more studies documenting the preg- Given the overlap between risk factors for hypertensive complications of pregnancy, and those for ESKD (e.g., BMI >35 kg/m 2 , CKD, diabetes mellitus), comparison of LKDs to the general population is unlikely to yield accurate attributable risk. 35,[52][53][54] To solve the conundrum of the adequate control group three studies included the pregnancy outcomes of the LKD both pre-and post-donation. Although the before-and-after design is useful for assessing short-term impacts, "threat to internal validity" may occur when assessing long-term impacts. Over a longer period, one is more likely to develop a confounding condition which may obscure the effects of an intervention, for example, an increased BMI postdonation. While the before/after design may appropriate risk more accurately than comparison to the general population, a more accurate comparator would be potential LKDs considered suitable for donation who did not proceed to donate. Accurate coding of these individuals in healthcare databases would allow for this group to be identified and their outcomes investigated.
The final limitation of all the studies reviewed was the absence of information on the LKD who did not have a post-donation pregnancy and the reasons why. Therefore, potential medical or psychosocial consequences of donation that may have precluded pregnancy in LKDs remain unknown.
As a result of the small sample sizes and inconsistent comparator groups, we were unable to perform a meta-analysis to provide a single estimate of effect. Our search for studies was comprehensive, albeit limited to English language studies. Furthermore, we do not strongly suspect publication bias as both negative and positive studies have been published.
Our second aim was to identify guidelines, consensus statements, and expert opinions which included the issue of pregnancy in LKD. While the guidelines offered albeit limited guidance on the acceptance of women of childbearing age as potential LKD, they were broadly consistent in stating that overall pregnancy after kidney-donation was of a risk similar to that of the general population and that women should be informed of this risk. They varied in their guidance regarding enquiring into prior pregnancy-induced complications, such as gestational diabetes and post-donation pregnancy care. 34

| Communicating risks
Lentine and Segev propose that informed consent must be based on informing potential LKD of their baseline (pre-donation) risk, the risk attributable to donation and their subsequent absolute (postdonation) risk of adverse pregnancy outcomes. 45 Baseline risk comprises demographic (e.g., age, ethnicity) and clinical (e.g., BMI, smoking status) characteristics, which in combination with the risk attributable to donation, can be used to calculate an absolute risk score for hypertensive complications of pregnancy.
Established perinatal registries could be used to calculate the baseline risk of hypertensive disorders of pregnancy. To be able to combine both the baseline risk and the risk attributable to donation to create the absolute risk score, linkage of perinatal registry data to a robust large dataset, such as a prospective LKD registry is required.
In an ideal scenario, the absolute risk of hypertensive pregnancy complications for the potential LKD would be derived by inputting baseline characteristics into a predicted risk calculator, such as www.trans plant models.com.
Additionally, potential LKDs should be informed of the long-term maternal consequences of pregnancy-induced hypertensive complications. In the general population, pre-eclampsia is associated with a greater than two-fold risk of death from cardiovascular disease, a 5-to 9-fold increased risk of ESKD and an increased risk of diabetes. [55][56][57][58] A recent study found that women with a pre-donation pregnancy  3. Risk prediction models for hypertensive complications of pregnancy in LKD should be developed by linking LKD and perinatal registry data. One should accept that these registries, in the short term, may lack some variables required to individualize risk, that is, ethnicity.
4. LKD registries should prospectively collect data on suitable LKDs that did not proceed to donation. This will improve the quality of control groups required to understand long-term LKD outcomes.

Qualitative research to understand prospective LKDs'
concerns surrounding post-donation pregnancy is needed.
6. The long-term health consequences of hypertensive pregnancies in women post-kidney donation should be established.
12% often did, 12% rarely did and 6% never discussed it. Only 20% of respondents were able to accurately answer questions on absolute versus relative risk for rare outcomes. 60 While this survey may be subject to the inherent biases of self-reports, it suggests there is a need to improve and standardize the communication of risk to potential LKDs. Box 1 lists our guidance on this topic, Box 2 lists our recommendations for future studies, and Box 3, provides a guide for practitioners counseling potential LKDs.

| CON CLUS ION
The increased risk of hypertensive disorders in pregnancy postdonation is evident in four retrospective studies, although all were limited to some extent by sample size and the lack of well-matched controls. Though the relative risk of pregnancy-related complications in LKD increases relative to the risk in the non-donor, the absolute risk remains very low. The risk of long-term complications from hypertensive disorders in pregnancy is minimal. 57,61 Perhaps, the combination of these two factors is the reason why LKD guidelines vary in the degree of attention paid to this issue.
The LKD of the future is likely to differ from the LKD of yester- year. As such more focus should be placed on better identifying and individualizing risk for LKD in the face of both older age and higher BMI at kidney-donation and subsequent pregnancy, and in non-white LKD. Only by doing so will consent to donation be truly informed.
For now, one should keep in mind that a potential LKD's personalized risk remains unknown, however, the consensus is that a LKD could be reassured that the absolute risk of post-donation pregnancy complications to mother and baby remains low.

D I SCLOS U R E
The authors of this manuscript have conflicts of interest to disclose

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.

R E FE R E N C E S BOX 3 A summary for practitioners counseling women on pregnancy post-living kidney donation
To date, 16