VP19.08: Late diagnosis of Kell alloimmunisation: a case report

Abstract

Objectives: To investigate if the standard fetal hemoglobin concentration curve developed by Nicolaides et al (1988) is suitable for prediction in a sample of Brazilian fetal population. Methods: Cross-sectional observational study that included 161 fetuses assisted at the Hospital of Clinics of Federal University of Minas Gerais (HC-UFMG) from January 2012 to December 2015. Considering ethical concerns about fetal blood sampling, our study was restricted to fetuses that would be submitted to cordocentesis for assistance reasons, wich is the case of prenatal karyotyping at HC-UFMG. Patients with singleton pregnancies between 18 and 36 weeks of gestation whose fetuses presented malformations demanding invasive prenatal genetic study were included. An aliquot of collected blood was reserved for hematological analysis, performed trough automated methodology. Absolute difference and variance between observed and expected hemoglobin based on Nicolaides et al formula were estimated. 48 fetal blood samples were available to final analysis once patients whose baseline condition could cause hematological changes were excluded. Simple linear regression was used to evaluate dispersion and compare haemoglobin levels. Paired t-test was applied to compare means. Pearson's, intraclass correlation and Bland-Altman diagram were obtained to assess measures agreement. Results: Hemoglobin levels were significantly more dispersed than expected by the standard curve, preventing the adjustment of a reliable regression model to estimate hemoglobin from our sample. Mathematic correlation tools showed non-agreement between expected and observed hemoglobin values. Conclusions: The curve proposed by Nicolaides et al does not fit as suitable model for prediction of fetal hemoglobin in the sample studied. Thus, it cannot be considered the best classification and treatment tool for anemic fetuses in our population.