| Autor: |
Elizabeth Wan, Daniela Iancu, Emma Ashton, Keith Siew, Barian Mohidin, Chih-Chien Sung, China Nagano, Detlef Bockenhauer, Shih-Hua Lin, Kandai Nozu, Stephen B Walsh |
| Rok vydání: |
2021 |
| Předmět: |
|
| Popis: |
BackgroundDistinguishing patients with the inherited salt-losing tubulopathies (SLT), Gitelman or Bartter syndrome (GS or BS) from wildtype (WT) patients who purge is difficult. We decided to identify clinical/biochemical characteristics which correctly classify SLT.Methods66 patients with possible SLT were recruited to a prospective observational cohort study at the UCL Renal Tubular Clinic (London). 31 datapoints were recorded on each patient. All patients were genotyped for pathogenic mutations in genes which cause SLT; 39 patients had pathogenic variants in genes causing SLT. We obtained similar datasets from cohorts in Taipei and Kobe; the combined dataset comprised 419 patients, 291 had genetically confirmed SLT. London and Taipei datasets were combined to train machine learning (ML) algorithms. These were then tested on the Kobe dataset to determine the best biochemical predictors of genetic confirmation of SLT.ResultsSingle biochemical variables (e.g. plasma renin) were significantly, but inconsistently different between SLT and WT, in the London and combined cohorts.A decision table algorithm using serum bicarbonate and urinary sodium excretion (FENa) achieved a classification accuracy of 74%. A simpler algorithm based on the FECl achieved a classification accuracy of 61%. This was superior to all of the single biochemical variables identified previously. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
