In newborns, intravascular hemolysis (the breakdown of red blood cells inside the blood vessels) can range from mild, as part of the physiological (normal) turnover of red blood cells, to severe in cases such as jaundice (an increase in bilirubin levels) Early biomarkers of haemolysis would improve neonatology (newborn) practice by identifying at-risk patients, particularly if the assay is simple, rapid, non-invasive and quantitative.

Our now-completed URICA trial on full-term male babies showed that the small cytoplasmic protein carbonic anhydrase I (CAI), found abundantly in red blood cells, was detected in 17 out of 26 urine samples collected once per recruited baby at the neonatology ward. CAI-positive samples were obtained from babies with levels of bilirubin that were rapidly rising or peaking above the threshold for phototherapy. CAI-negative urine was obtained when either bilirubin did not reach phototherapy (a light treatment used for excessive jaudice) threshold, or after it had recovered from its peak. On four occasions, the cause of CAI-positive urine was undetermined. Since CAI is normally absent from urine, a positive signal is indicative of intravascular hemolysis and confirms that CAI crossed the glomerular barrier (a barrier within the kidneys that filters large molecules). However, the quantitative power of urinary CAI to predict and estimate an impending haemolytic crisis requires a new longitudinal study, which is the objective of the URICA-II trial.

The URICA-II trial would recruit 30 full term newborn infants delivered at the Evelina London Children's Hospital. The babies recruited would be expected to stay in the hospital for at least 5 days due to treatment for jaundice, infection or some other condition.

Participants will have daily non-invasive (bag) urine samples collected and daily transcutaneous (skin) bilirubin levels recorded upto 10 days.

The study will last upto 2 years.

In newborns, intravascular hemolysis can range from mild, as part of the physiological turnover of red blood cells, to severe in cases such as jaundice or ABO incompatibility. Early biomarkers of haemolysis would improve neonatology practice by identifying at-risk patients, particularly if the assay is simple, rapid, non-invasive and quantitative. Our now-completed URICA trial on full-term male babies (Trial registration CPMS 19576) showed that the small cytoplasmic protein carbonic anhydrase I (CAI), found abundantly in red blood cells, was detected (by immunoassay and mass-spec), in 17 out of 26 urine samples collected once per recruited baby at the neonatology ward. CAI-positive samples were obtained from babies with levels of bilirubin that are rapidly rising or peaking above the threshold for phototherapy. CAI-negative urine was obtained when either bilirubin did not reach phototherapy threshold, or after it had recovered from its peak. On four occasions, the cause of CAI-positive urine was undetermined (https://doi.org/10.1093/jalm/jfaa051). Since CAI is normally absent from urine, a positive signal is indicative of intravascular hemolysis and confirms that CAI crossed the glomerular barrier. However, the quantitative power of urinary CAI to predict and estimate an impending haemolytic crisis requires a new longitudinal study, which is the objective of the URICA-II trial.

Clinical rationale: The health of neonates must be monitored regularly for one of many rapidly progressing and potentially life-threatening conditions that implicate intravascular hemolysis. These include sepsis (e.g. Gram-positive bacteria), protozoan infection (e.g. malaria), alloimmunity (ABO and Rhesus incompatibility), genetic traits (e.g. sickle cell, glucose-6-phosphate dehydrogenase [G6PD] deficiency) or electrolyte imbalance. Some of these conditions (e.g. RhD, G6PD, sickle trait) are particularly prevalent in minority ethnic groups [1-4]. In advanced healthcare facilities, tests for haemolysis involve taking blood samples for microscopy analyses (e.g. cell count and morphology) or biochemical assays (e.g. for unconjugated bilirubin or lactate dehydrogenase). Other resource-intensive tests are available to diagnose specific causes of haemolysis, such autoimmune haemolytic anaemia with Coombs test or G6PD deficiency by redox assays. These tests require trained personnel, advanced laboratory equipment and high standards of hygiene. Simpler, cheaper and faster tests on novel biomarkers of disease, adapted for point-of-care, minimal-laboratory conditions, would be medically and economically justified. This technology would also benefit the developing world, where hemolysis can be more common.

Scientific background: The soluble enzyme carbonic anhydrase I (CAI) is, after haemoglobin isoforms, the most abundant protein in the cytoplasm of red blood cells [5]. Plasma normally contains only trace amounts of CAI but levels will increase upon haemolysis [6]. Based on its small size (29 kDa and Stokes radius 25 Å [7-9] CAI is predicted to filter across the glomerular barrier and appear in urine [10]. Quantitative immuno-techniques can detect CAI even at 1:10,000 dilution. Urine from healthy adults is normally devoid of CAI, producing a very low background signal. Urinary CAI excretion is expected to be proportional to the extent of hemolysis, and therefore serve as a quantitative biomarker that can be assayed non-invasively.

Objectives of the (first) URICA trial: The objective of the now-completed URICA trial (URInary excretion of Carbonic Anhydrase; trial registration CPMS 19576) was to test for the presence of CA isoforms in the urine of full term neonates admitted to Evelina's Children Hospital. The cohort included controls who were not suspected of undergoing hemolysis (beyond the mild physiological form) as well as babies with ABO incompatibility, anaemia and undergoing treatment for jaundice. The URICA trial obtained ethical approval on May 12th, 2015 (REC: 15/NS/0042). The collection and analysis of samples was completed by June 2017. Measurements on urine comprised of western blotting for CAI (a sensitive but semi-quantitative method), ELISA for CAI (a quantitative method for determining CAI levels), mass-spec for obtaining the proteome of urine, plus measurements of total protein and creatinine (to assess evidence for renal failure as an exclusion criterion). In addition, clinical data were collected about gestation, and blood haemoglobin, serum or transcutaneous bilirubin and C-reactive protein at several time points, where possible, to test for anaemia, jaundice and infection, respectively. 26 babies were recruited.

Outcomes of the URICA trial: Western blotting showed no detectable CAI signal in 9 samples (indistinguishable from background-signal in adult urine) and a range of positive signals in 17 samples. These results were supported by ELISA measurements. Two samples had elevated protein/creatinine levels, possibly indicating renal damage yet these did not produce CAI-positive urine. Mass spec analysis showed that CAI level did not correlate with any other red blood cell protein, indicating strongly that the hemolysis had occurred inside vessels and not in the urinary tract (where it would also lead to signals from e.g. haemoglobin or anion exchanger 1). When compared to CAI-negative urine, the most over-abundant protein in CAI-positive samples was, in fact, CAI, arguing for the excellent resolving power of this biomarker. When considering the time point at which urine was collected, it is possible to explain the CAI signal relative to bilirubin measurements as follows:

• Six babies with subthreshold bilirubin levels (NHS guidelines for phototherapy) produced CAI-negative urine;
• In sixteen babies with rapidly and significantly rising bilirubin levels (14 of which received phototherapy), the CAI signal was strongest in urine samples that were collected nearer to the bilirubin-peak, with a tendency to suggest that the CAI spike preceded the bilirubin peak;
• Four babies produced urine with a very strong CAI signal but did not manifest elevated bilirubin, arguing that a distinct mechanism is producing hemolysis.

Hypotheses: We propose that urinary CAI is a biomarker of hemolysis, and we hypothesise that:

1. Urinary CAI signal can be stratified to relate to the underlying haemolytic state:

   • undetectable: indicating no evidence for hemolysis;
   • intermediate: levels that indicate hemolysis in jaundiced babies;
   • higher threshold: indicating an acute haemolytic crisis which may not invoke a rise in bilirubin.

2. In the intermediate range, a positive urinary CAI signal precedes the peak of bilirubin and is an early indicator of jaundice.

5 Trial objectives and purpose

The time course of bilirubin is currently used to gauge neonatal health, but we propose that urinary CAI monitoring would provide an early-warning sign for identifying at-risk patients and thus allowing neonatal wards to better rationalize their resources. To establish the utility of this biomarker, we propose to undertake a longitudinal study on a cohort of full-term babies to correlate the time courses of urinary CAI with bilirubin (to test hypothesis 2), and to obtain clinical information that may determine the cause and extent of hemolysis (to inform threshold levels for hypothesis 1).

Research plan: There are 9,000 births a year at St Thomas'. Of these, ~1000 babies are admitted to the Neonatal Intensive Care Unit (NICU) each year; approximately half of these babies are born at term. Some ~300 will have jaundice needing phototherapy and ~10-20 will have severe haemolysis (typically ABO incompatibility). The aim is to recruit 30 full-term babies, over a period no longer than 24 months, who are expected to remain on the neonatal unit for at least five days and up to 10 days during which urine samples will be obtained as regularly as possible (once a day; if not possible, then at least once every two days), time-stamped and stored in a dedicated freezer for off-line analysis in Oxford, once recruitment is completed. When feasible to perform close to the point of collection, a small volume of urine (no more than four droplets) will be pipetted onto a lateral flow device (LFD), provided as part of a kit, to obtain an instantaneous readout of CAI immunoreactivity. The LFD be labelled with the matching urine identifier and recorded (positive/negative/void). If not feasible near the point of collection, this step will be performed later, in Oxford, on thawed samples. Within 6 hours of the urine sample, a non-invasive transcutaneous bilirubin level will be obtained and recorded. These data will be complemented with the following information, obtained as part of the normal clinical practice unrelated to the objectives of the trial.

• Serum bilirubin where routinely collected
• Blood haemoglobin and full blood count
• C-reactive protein (CRP) and renal function
• Gestational age;
• Duration and timing of phototherapy
• Birthweight
• Sex
• Result of direct Coombs test;
• Reason for being admitted to the NICU.

Urine samples will be analysed in Oxford in terms of:

• CAI immunoreactivity on western blot, a highly sensitive semi-quantitative method;
• CAI immunoreactivity by ELISA, a quantitative method;
• Total protein and creatinine as markers of kidney damage;
• Osmolarity, to determine diuretic state. All samples will be anonymised with an alphanumerical code and linked to clinical data. Babies from whom less than 4 urine samples are obtained will be excluded from the study on the basis that such datasets would not have the resolving power to test the hypotheses. Recruitment will complete when 30 babies with at least 4 samples have been recruited. Babies with chromosomal abnormalities will also be excluded from the study.

Inclusion Criteria:

Greater than 36 weeks gestation infants admitted to the neonatal unit and expected to be an in-patient for at least 5 days

Exclusion Criteria:

Babies with chromosomal abnormalities

Greater than 36 week gestation babies admitted to the neonatal unit

No plan to share IPD outside the study team