Pregnancy and kidney disease: It is time for the birth of prospective registries

Abstract

Pregnancies in women with chronic kidney disease (CKD), on dialysis, and those with kidney transplants are rare events and often medically challenging. They require careful preconception planning and heightened antenatal surveillance to achieve the best possible maternal and foetal outcomes. Knowledge about pregnancies in the setting of significant maternal kidney disease has grown substantially in the last two decades. There has been a definite paradigm shift away from the fear of pregnancy (or at worst, the forbidding of it) to a more permissive attitude with patient-centred, shared decision-making that explicitly incorporates a woman s preferences, appetite for risk and autonomy. Pregnancy is a key goal and aspiration for many women with CKD. Women with advanced CKD who wish to have a family face enormously complex decisions about the timing of pregnancy. Do they attempt pregnancy now, when the ‘stress test’ of pregnancy may accelerate them towards kidney failure? Do they wait until after transplantation? What are the outcomes at each stage of kidney disease? Counselling women with kidney disease requires robust data upon which to base decisions. While core pregnancy outcomes in women with kidney transplants are reasonably well-defined thanks to excellent data from renal registries, more granular obstetric or perinatal data are not usually captured. The cohort of women without kidney failure is challenging to study for several reasons: firstly, there is a lack of registries focused on earlier CKD stages, and secondly, there is very poor capture of CKD and CKD severity within population-level hospital, obstetric or perinatal datasets, either via diagnostic coding or actual biochemical data. In this issue of Nephrology, Chewcharat et al. add substantially to existing knowledge and further define pregnancy outcomes in women with moderate–severe CKD and kidney transplants. The authors analysed a large US database with over 5 million hospitalisations for births during 2009–2014 to explore the pregnancy outcomes in women with kidney transplants and CKD Stages 3–5 (based on ICD-9 Diagnostic coding data) and compared them to women with no coding for kidney disease. A particular strength of this study was the cohort size and consequent ability to carefully identify the CKD cohort, with appropriately adjusted analyses and matching for comorbidity and maternal demographics. This study confirms that transplanted women have a substantially increased risk of critical adverse events with maternal and foetal morbidity – acute kidney injury (AKI), hypertensive disorders of pregnancy, preterm delivery, foetal growth restriction, and Caesarean delivery. Women with coded Stage 3–5 CKD and unspecified CKD had similar adjusted odds compared to transplanted women for most events, except AKI, which was higher in the CKD cohort. When only women with coded CKD Stage 3–5 were included in the analyses, they had a longer hospital stay and higher hospitalization costs, higher risk of AKI, pre-eclampsia, and foetal death or stillbirth compared to transplanted women. Chewcharat et al. suggest that patients with CKD Stage 3–5 should delay pregnancy until after kidney transplantation as a means of improving outcomes. Such a delay may be feasible for some, but many women may not reach this point until after their childbearing window has closed, considering the time to reach kidney failure, waiting time for transplant and time to wait post-transplant until pregnancy is deemed safe. These findings may simply indicate that transplanted women who achieved pregnancy had better kidney function and health and may have had planned pregnancies, compared to women with CKD Stage 3–5 where kidney function (by definition) was poor, and pregnancies may have been unplanned. This remains entirely speculative in the absence of supporting biochemical and clinical data. Despite the authors best efforts, the study has limitations that are inherent with diagnostic coding data, and conclusions must be drawn with an understanding of the absent information and potential confounders within the dataset. CKD stage was based on diagnostic coding data, and how this was determined is unknown. Underreporting of CKD in the overall cohort is likely; women with unspecified CKD may have had CKD Stage 3–5; the CKD Stage 3–5 cohort may have reflected women with more overt and severe disease. Women with unspecified CKD stage were excluded – this approach is appropriate but excludes a large cohort. Furthermore, transplant details, including transplant-pregnancy interval and function, medications including immunosuppression, and data on non-hospitalized early pregnancy losses, were not available. This study exemplifies the challenges for obstetric nephrology research, given kidney function before and during pregnancy, renal pathology, comorbidities, drugs and disease history are critical factors influencing pregnancy outcomes. There are complexities and significant limitations with current hospital and perinatal datasets (that are not designed to capture CKD well) and renal registries (that are not designed to capture CKD or pregnancy data well). These limitations provide a clear mandate for developing purpose-built data systems to advance knowledge about pregnancy in women with CKD. Prospective data collection that includes essential elements required to understand risk stratification and both early and late pregnancy maternal-foetal outcomes is the logical next step. Sharing such data across national and international Received: 15 August 2021 Accepted: 23 August 2021