Congenital anomalies of the kidney and urinary tract, collectively known as CAKUT, are among the most common structural abnormalities identified in newborns. They range from relatively mild changes in the shape or position of the kidneys to severe malformations that disrupt urine production, drainage, or the development of both organs. For critically ill infants, the difference between a condition that can be managed over time and one that threatens life may become clear within hours of birth. A new study introduces a clinical risk prediction tool designed to help doctors estimate that danger earlier: the “CAKUT calculator.”
Published in the Journal of Perinatology, the study by Rumpel, Zaniletti, Stoops and colleagues focuses on neonates with CAKUT who are admitted to level IV neonatal intensive care units, the highest level of intensive care available for newborns. These units care for infants requiring advanced respiratory support, complex surgery, dialysis, continuous monitoring and other highly specialized treatments. The researchers set out to predict in-hospital mortality using perinatal factors—information available around the time of birth—rather than relying solely on complications that emerge later during a hospital stay.
The clinical challenge is substantial because CAKUT does not represent a single disease. It is an umbrella term covering kidney agenesis, renal dysplasia, obstructive abnormalities, urinary tract malformations and other developmental defects. Some infants have one abnormal kidney and remain relatively stable, while others have bilateral disease that severely limits renal function. Kidney failure can affect fluid balance, blood pressure, electrolyte concentrations and acid-base chemistry. In the newborn period, these disturbances may develop alongside respiratory failure, infection, cardiovascular instability or the need for urgent surgery, making early risk assessment particularly difficult.
The proposed calculator translates several perinatal characteristics into an individualized estimate of mortality risk. In medical prediction modeling, such a tool typically combines patient-level variables through a statistical algorithm. Each factor contributes a specific amount to the final prediction, allowing the model to transform a complex clinical profile into an interpretable probability. Depending on the variables selected by the investigators, the calculation may incorporate features such as gestational age, birth weight, sex, prenatal diagnosis, congenital conditions and the severity of illness at admission. The essential principle is that the estimate is generated from information available early enough to support decision-making.
A prediction model can be especially valuable in CAKUT because the first clinical decisions often must be made before the full anatomy and long-term prognosis are known. Neonatologists may need to determine how aggressively to stabilize an infant, whether transfer to a highly specialized center is necessary, how urgently nephrology or urology should be involved, and which families require immediate counseling. A calibrated risk estimate cannot replace clinical judgment, but it can provide a consistent framework for interpreting multiple risk factors at once. It may also help clinicians identify infants who appear more vulnerable than their initial physical examination suggests.
The study’s focus on level IV NICUs is important for another reason: these hospitals generally receive the most medically complex newborns, including infants referred from other facilities because of severe malformations or organ failure. Outcomes in such settings can be influenced by referral patterns, surgical capacity, access to pediatric nephrology and the availability of renal replacement therapies. A model built specifically for this population is therefore intended to reflect the realities of high-acuity neonatal care rather than the broader population of infants with CAKUT, many of whom never require intensive care.
An interactive calculator could make the research more accessible at the bedside. Instead of asking clinicians to interpret a table or manually apply regression coefficients, an online interface can allow them to enter a newborn’s characteristics and receive a calculated estimate almost immediately. Such tools can also reduce arithmetic errors and make the underlying model easier to use across hospitals. However, the reliability of any calculator depends on how well its predictions are validated in different populations. A model developed from one healthcare network or a particular group of intensive care patients may perform differently in another region, among different racial and ethnic populations or in hospitals with different treatment resources.
The researchers’ work also highlights the growing role of clinical prediction models in neonatal medicine. Modern neonatal care produces large quantities of structured information, from prenatal imaging and delivery-room events to laboratory measurements and organ-support requirements. Statistical models can identify patterns that are difficult to recognize when dozens of variables must be considered simultaneously. Yet technical performance is only one part of clinical usefulness. A successful tool must also be transparent, understandable, regularly updated and tested for calibration, meaning that predicted risks should correspond closely to outcomes observed in real patients.
For families, mortality estimates are emotionally sensitive and must be communicated with care. A probability is not a destiny, and no calculator can capture every factor influencing an individual infant’s course. Treatment decisions may change rapidly as kidney function evolves, surgery becomes possible or an infant responds to intensive care. The CAKUT calculator is best understood as an aid to structured assessment and shared decision-making, not as an automated verdict. Its greatest potential may lie in helping medical teams recognize risk early, coordinate multidisciplinary care and explain uncertainty more clearly during some of the most difficult conversations in neonatal medicine.
By targeting a population in which anatomy, physiology and critical illness intersect, the new tool aims to bring greater precision to the first days of life for infants with CAKUT. If externally validated and integrated responsibly into neonatal practice, it could support earlier consultation, improve communication between referring hospitals and tertiary centers, and provide researchers with a common framework for comparing outcomes. The calculator does not eliminate the complexity of congenital kidney disease, but it represents a step toward making that complexity more measurable—and potentially more manageable—when every hour can matter.
Subject of Research: Prediction of in-hospital mortality in neonates with congenital anomalies of the kidney and urinary tract admitted to level IV neonatal intensive care units.
Article Title: CAKUT calculator: a clinical risk prediction tool for in-hospital mortality in neonates with congenital anomalies of the kidney and urinary tract admitted to level IV NICUs
Article References: Rumpel, J.A., Zaniletti, I., Stoops, C. et al. “CAKUT calculator: a clinical risk prediction tool for in-hospital mortality in neonates with congenital anomalies of the kidney and urinary tract admitted to level IV NICUs.” Journal of Perinatology (2026). https://doi.org/10.1038/s41372-026-02823-6
Image Credits: AI Generated
DOI: 10.1038/s41372-026-02823-6
Keywords: CAKUT, congenital kidney anomalies, neonates, neonatal intensive care, in-hospital mortality, clinical risk prediction, NICU, kidney disease, neonatal medicine, risk calculator
Tags: CAKUT in neonatal intensive careCAKUT risk assessment toolclinical decision support for newborns with CAKUTcongenital kidney and urinary tract anomaliesearly detection of CAKUT severitylevel IV neonatal intensive care unitsmanagement of congenital renal anomaliesneonatal in-hospital mortality predictionneonatal intensive care mortality risk factorsneonatal kidney malformationsperinatal factors in neonatal prognosispredictive modeling in neonatal care
